(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Metallurgy Of Copper"

TIGHT BINDING BOOK 



LLJ < 



OU 



168138>3 



OSMANIA UNIVERSITY LIBRARY 

. / ' i > H ft n -, c ^ - x, 

(.'-all No <-> v f T -* Accrssion No ^ ^ x '-^ 

^i -* . 
\nthor 

: . !' f i'. v-1. V "f' 11 / 

'I his book should br n turiK'd on or brforr the d tie List m.irkf d bcl< \ 



METALLURGY 

of 
COPPER 



RY 
JOSEPH NEWTON 

Assistant Professor of Mctsillurgy 

I nin'rsit^ of Idaho 

Mos( (m , I (I (i ho 



CURTIS L. WILSON 

/Von, Missouri ^( hoot of Mines and Metallurgy 

I Orrncr /Vo/rssor <tf \lctallnrgy 

Montana ^<hool of \1ines 



NK\V YORK 
\\ILKY *Jv SOi\ T S, 4NC. 

LOM>ON: Cll \PM\N vK II VLL. LIMITED 



COPTHIOHT, 1942 

BT 

JOSEPH NEWTON AND CURTIS L WILSON 



All Rights Resented 

Th\9 book or any part thereof must not 
be reproduced in any form without 
the written permission of the publisher. 



PREFACE 

The aim of this book is to present a discussion of the various methods 

employed in winning copper from its ores and in refining the metal to 

commercial grade. Examples of modern practice are included to 

illustrate the application of these methods, but no attempt has been 

ma-le to compile a complete and exhaustive treatise on the practice all 

over the world. Such a treatise might well require several volumes. 

'onfining the discussion largely to the extraction and refining of 

r, it has been possible to touch only lightly on several related 

cts because of space limitation?. The chapter on ore dressing is 

m. % ly a summary to indicate the methods used in dressing copper ores 

and the nature of the resulting concentrates. It was not possible to 

consider the subject of copper alloys in any great detail. 

An attempt has been made to give credit at the proper place for all 
material used in the book. The authors extend their thanks to the 
various mining, smelting, refining, and manufacturing companies, and 
to the publishing companies for their kind and willing cooperation. 

JOSEPH NEWTON 
CURTIS L. WILSON 

June, 1942 



CONTENTS 

CHAPTER PAGE 

I. FROM ORE TO CONCENTRATE 1 

II. THE EXTRACTION OF COPPER FROM ITS ORES .... 32 

III. ROASTING 50 

IV. SMELTING 76 

V. CONVERTING 162 

VI. FIRE REFINING 188 

VII. SMOKE AND OASES 226 

VIII. ELECTROLYTIC REFINING 250 

IX. HVDROME'I U.LVRGY . 303 

X. PROPERTIED OF COPPER 379 

XI. THE USES OF COPPER 396 

XII. PRODUCTION OF COPPER 430 

BIBLIOGR \PIIY . . 499 

NAME INDEX 501 

SUBJECT INDEX 503 



CHAPTER I 

FROM ORE TO CONCENTRATE 
THE IMPORTANCE OF COPPER 

From the beginning of recorded history until the end of the medieval 
period, copper was the world's most useful metal. Its use marked the 
transitory step in the progress of civilization from the Stone Age to the 
Metal Age. Although gold, owing to its sparkling yellow color, its 
high luster, its resistance to corrosion and tarnish, and its occurrence in 
the free or elemental state in nature, was unquestionably the first metal 
to attract the attention of man, and although in certain localities iron, 1 
in the form of meteorites or even obtained by the reduction of the oxide 
with charcoal, may have been used before copper, nevertheless every 
ancient metal culture was actually introduced by the use of copper. 2 
In the form of pure metal, fashioned first by the crude hammering of 
masses of native copper and later by melting, and in the form of 
bronze, obtained by smelting mixed tin and copper ores, it was employed 
originally for ornaments and statues and then as tools, domestic 
utensils, implements of war, and for every purpose in which its 
strength, hardness, and toughness proved its superiority to stone, wood, 
and other materials. 

When the methods of producing iron evolved from the direct 
processes through the cast iron period to puddling, cementation, and 
the crucible process, iron and steel usurped copper's position of first 
importance; and with the advent of the Bessemer and open-hearth 
processes, ferrous materials attained such ease of large-scale production 
and such widespread use that they almost eclipsed copper The age of 
electricity, however, introduced new requirements for materials to be 
used in the generation and transmission of electrical energy; copper 
immediately entered its rejuvenation 3 and assumed firs* place in 
importance in the electrical field and second in general utility in our 
present-day civilization. 

1 Howe, H M , The Metallography of Steel and Cast Iron, p. 4, McGraw-Hill 
Book Co , New York, 1916. 

Rickard, T. A , The Early Use of Metals: Jour. List. Metals, Vol. 43, p. 297, 
1930. 

3 Davis, Watson, The Story of Copper, p. 58, D. Appleton-Century, New York, 
1924. 

1 



2 FROM ORE TO CONCENTRATE 

Next to iron, then, copper is the world's most important metal. It is 
important for three primary reasons: (1) because of its abundance, 
assuring a supply which will make possible its continued use in large 
quantities; (2) because of its high electrical conductivity, surpassed 
only by one other substance, silver, which is not abundant enough, 
cheap enough, nor strong enough to acquire a utilitarian role similar to 
that of copper in the electrical field; and (3) because of the important 
alloys which it forms. Chief among these are the brasses and bronzes. 
Brass is our most widely used non-ferrous alloy and therefore ranks 
second only to steel in all the alloys in use today. 

COPPER ORE MINERALS 

Copper is an important component of many minerals, a surprisingly 
large number of which are today ore minerals, that is, of industrial 
importance. These minerals are listed below in Table 1, together 
with their supposed chemical formulas and their theoretical compo- 
sitions. By " theoretical " is meant that composition which corresponds 
to the supposed chemical formula, for although according to a common 
definition " a mineral is a naturally occurring substance of definite 
and uniform chemical composition with corresponding characteristic 
physical properties," nevertheless the chemical composition of many 
of the copper minerals does vary within limits For this reason 
differing chemical formulas have been assigned to them from time 
to time, derived almost always from the chemical analysis alone. 
Bornite, for example, is written Cu :i Fc$ a (SCuoS-FcoSs) and 
Cu 5 FeS 4 (5Cu 2 S*Fe 2 So), the former containing 55.6 per cent copper 
and the latter 63 3 per cent. 

Minerals might be regarded in the same light as alloys, as being com- 
posed of mixtures or solid solutions of various components, which in 
the case of minerals are definite chemical compounds and elements. 
Chalcopynte, to use a common example, is usually designated by the 
chemical formula CuFeS 2 , which enlightens one merely w r ith respect 
to the chemical composition but tells nothing about the structure or 
constitution. It can also be written Cu 2 S*Fe 2 S ;i , showing that it is an 
association of the two chemical compounds Cu 2 S and Fe 2 S 3 . There 
is some doubt as to the existence of such a compound as Fe 2 S, H ; it 
might be the well-recognized chemical compound FeS with some sulfur 
in solid solution, as in the case of pyrrhotitc. The laws of heterogen- 
eous equilibrium will eventually establish the true constitution of 
minerals and definitely prove whether chalcopyrite, to use the same 
example again, is the cuprous salt of the hypothetical acid HFeS 2 or 
whether it is to be regarded as analogous to a pseudo-binary alloy and 



COPPER ORE MINERALS 



therefore composed of a definite chemical compound Cu 2 S'Fe 2 S3, 
formed by the combination of the two components Cu 2 S and Fe 2 S 3 , or 
finally whether it is to be regarded as a ternary system with the three 
components Cu 2 S, FeS, and S. Investigations of such problems can 
be made only at great expense and with much patient effort, for al- 
though the study of chalcopyrite in the Cu 2 S-FeS-S ternary system 
would be confined to only a portion of the entire Cu-Fe-S system, 
nevertheless, the variables would have to include not only composition 
and temperature but also pressure. Because the conditions under 
which minerals are formed are so difficult to duplicate, we cannot say 
with authority whether or not complex arrangements actually exist. 
In all probability they do not 

The formulas given in the table are for the most part only approxima- 

TABLE 1 
COPPER ORE MINERALS 



Mineral 


Formula 


Composition 
(per rent) 


Cu 


Fe 


S 


As 


Sb 


Native : 














Native Copper 


Cu 


99 9 











Sulfide: 














* Chalcocite 


Cu 2 S 


79 9 




20 1 






. Covelhte 


CuS 


66.5 




33.5 




.... 


Chalcopyrite 


CuFeS2 or Cu2$'Fe2S 3 


34.6 


30 5 


34 9 






Bornite a 


Cu 3 FeS 3 or 3Cu 2 S Fe 2 S 3 


55 6 


16 3 


28 1 




.... 


Enargite 


Cu 3 AsS 4 or 3Cu 2 S As 2 S 5 


48 4 




32 6 


19 


.... 


Famatimte 


Cu 3 SbS 4 or 3Cu 2 S Sb 2 S 5 


43 3 




29 1 




27 6 


. Tetrahednte 6 


Cu 3 ShS 3 or3Cu 2 8Sb 2 S (i 


46 7 




23 5 




29.8 


Tennantite c 


Cu 3 AsSs or 3CU2S As2S 3 


52.7 




26 6 


20 7 




Oxidized . 














Cuprite 


Cu 2 O 


88 8 










Tenonte 














(rnel aconite ) 


CuO 


79.9 










IVtalachite 


CuCO 3 -Cu(OH) 2 


57 5 










Azunte 


2CuCO 3 Cu(OH) 2 


55 3 










Chrysocolla 


CuSiO 3 2H 2 O 


36 2 








.... 


Chalranthite 


CuSO 4 5H 2 O 


25 5 










Brochantite 


CuSO 4 3Cu(OH) 2 


56.2 










Atacamite 


CuCl 2 3Cu(OH) 2 


59 5 










Krohnkite 


C 1 uSO 4 Na 2 SO 4 3Cu(OH) 2 


42 8 















a Bornite also written Cu & FeS4 or 5Cxi2 

6 Tetrahednte also written Cu 8 Sb 2 S 7 or 4Cu 2 S-Sb2Sa 

c Tennantite also written 01^8287 or 



4 FROM ORE TO CONCENTRATE 

tions; the percentage composition refers to the pure minerals and not 
to the ores. The copper in many minerals may be replaced by lead 
or some other metal, and the arsenic and antimony by each other. 

As the source of the world's total production of copper, chalcocite 
represents approximately one-half, chalcopyrite one-quarter, enargite 
3 per cent, other sulfides 1 per cent, native copper 6 or 7 per cent, and 
the oxidized copper minerals some 15 per cent. 

Native copper occurs in most of the principal copper deposits of the 
world, but usually in small quantities. It has been found in 27 states 
of the United States, in Bolivia, Chile, Australia, and elsewhere. 
The deposits of the Lake Superior district are the only ones of economic 
importance, however. The metal is very pure, containing from 98 to 
99 92 per cent copper, with small amounts of silver which are mechan- 
ically enclosed and not alloyed. From the standpoint of the genesis 
of ore deposits this fact is important, for had the metals been de- 
posited from a molten magma they would exist as an alloy. Some 
iron, arsenic, nickel, bismuth, and mercury may also be present but 
strangely no gold. 

Chalcocite is a steel-gray mineral with a metallic luster, often tar- 
nishing to a dull blue or green. It crystallizes in the orthorhombic 
system, but distinct crystals are rare, the occurrence being commonly 
massive. Veins more than 20 feet across have been found in Butte 
and in Alaska. Its cleavage is indistinct and its fracture conchoidal. 

Covellite, the cupric sulfide, is less stable chemically than chalcocite, 
the cuprous sulfide. It has a beautiful deep indigo blue color which 
upon being moistened turns to a characteristic, easily recognizable 
purple. It is a relatively rare ore mineral except at Butte, Montana, 
where it occurs massive in some of the mines. 

Chalcopyrite is geographically the most widely distributed copper 
mineral, occurring in practically every copper field in the world. It 
is not the prevailing mineral of the greatest producing mines, however, 
and therefore ranks after chalcocite as a source of copper. It has a 
brass-yellow color, metallic luster, a greenish-black streak, and occurs 
usually in compact masses, although occasionally in crystals of the 
tetragonal system. It is considered the primary ore of copper along 
with bornite and cupriferous pyrite, and from these all other (second- 
ary) copper minerals were generated. The theoretical composition 
is about 34.6 per cent copper, but the copper content may be as little 
as 2 per cent or less. Even such low-grade deposits can be smelted 
profitably under favorable conditions. 

Bornite is another ore mineral in which the copper content varies, 
as has already been mentioned. It is fairly common but occurs 



COPPER ORE MINERALS 5 

usually in subordinate amounts. The freshly broken surface exhibits 
a copper-red to bluish-brown color (" horseflesh ore "), which tarnishes 
to variegated blues and purples, from which it has likewise derived 
the name " peacock ore." Other complex sulfidcs, less important than 
bornite, are chalmersite CuFe 2 S 3 , and cubamte CuFe 2 S 4 . 

Enargite, the sulfarsenate of copper, is a relatively rare mineral 
except in Butte, where it occurs in such large quantities that it has 
become the source of about 3 per cent of the world's copper production. 
It is a brittle, grayish-black mineral containing 19 per cent arsenic and 
as such is an important raw material for the byproduction of arsenic 
trioxide. When the arsenic is replaced by antimony, the sulfanti- 
monate famatinite results. Where the Cu 2 S is associated with anti- 
mony trisulfide Sb 2 S 3 , the sulfantimonite tetrahedrite, or gray copper 
ore, is found. The copper is often replaced by iron, zinc, mercury, or 
lead, as in bournonite 3(Pb,Cu 2 )S-Sb 2 S 3 . Silver is usually present in 
tetrahedrite, making it an ore of silver as well as of copper. Tetra- 
hedrite is the principal ore mineral at the Sunshine and other silver 
mines in the Coeur d'Alene mining district in Idaho. The sulfarsenite 
corresponding to tetrahedrite is called tennantite. 

Cuprite, the cuprous oxide, Cu 2 0, occurs in the upper zones of most 
oxidized copper ore deposits and, in early developments, was an 
important ore mineral. It is usually some shade of red or brown and 
in translucent crystals shows a ruby red, from which it has derived 
the name " ruby ore." The cupric oxide CuO is black and is known as 
tenorite or melaconite. 

Malachite is the most abundant oxidized ore of copper, occurring 
usually in copper veins which he in limestone. It has a beautiful green 
color, and when found in large solid masses, many of which are 
artistically marked, it is valuable not only as an ore of copper but 
also as a semi-precious stone, used for jewelry, table tops, vases, and 
other works of art. The pure mineral contains 57% per cent copper, but 
because of its high coloring power and solubility it often stains and in- 
crusts large areas of worthless rock, disguising it as valuable mineral. 

Azurite, like malachite, is a basic copper carbonate but is less widely 
distributed. It possesses an intensely azure blue color from which it 
gets its name. When associated in alternating concentric rings with 
malachite, the contrast of colors is striking. 

Chrysocolla is the only important silicate of copper. It likewise has 
a green to greenish-blue color but is non-en stalline and earthy in 
appearance. It occurs in commercial quantities in Arizona, Chile, 
and the Belgian Congo. Other silicates of copper, such as dioptase, 
cornuite, plancheite, shattuckite and bisbeeite, are rare. 



6 FROM ORE TO CONCENTRATE 

Chalcanthite is an oxidation product of minerals containing copper 
sulfide. It is found dissolved in mine waters and crystallizes in the 
form of stalactites or incrustations. 

Brochantite, the basic copper sulfate, is not of major importance in 
the United States but it occurs as the principal ore mineral in the 
oxidized zone at Chuquicamata, Chile. The basic chloride atacamite 
likewise occurs massive in Chile and Bolivia. Krohnkite, found in the 
upper zone at Chuquicamata, is the basic sulfate of sodium and copper. 

COPPER ORES 

Copper ores are widely distributed throughout the world, occurring 
in every continent and in almost every country. They are furthermore 
found in practically every type of ore deposit and are associated, in 
one place or another, with every metallic and rock-forming mineral. 
This distribution, though wide, is not uniform, so the present world 
production comes mainly from certain definite, limited localities. 

There are four major known sources of supply in the world at present. 
In the order of their importance as gaged by past production they are 

(1) the Rocky Mountain and Great Basin area of the United States; 

(2) the west slope bf the lAndes in Peru and Chile; (3) the central plateau 
of Africa in the Belgian Congo and Northern Rhodesia; (4) the pre- 
Cambrian shield area of central Canada and its extension into northern 
Michigan. These areas contain about 95 per cent of the total known 
reserves. 4 

Types of Copper Ores. In the metallurgical treatments required for 
the winning of metallic copper we may make the following rough 
classification of copper ores. 

1. Sulfide Ores: 

a. High-grade, direct-smelting ores. 

b. Medium-grade ores which must be concentrated. 

c. Low-grade ores which require concentration and must be mined and milled 

on a large-scale, low-cost basis. 

d. Pyritic ores. 

2. Oxidized Ores : 

a. High-grade or medium-grade ores which can be smelted to " black copper " 

by reduction smelting, mixed with sulfide ore or concentrate for matte 
smelting, or leached. 

b. Low-grade ores which are treated by leaching. 

3. Native Copper Ore. 

Under Ic and 2b appears the group of copper ore deposits which is 
the most economically important of all, the porphyry coppers. Twelve 
of these immense deposits are now being exploited nine in south- 

4 Notman, Arthur, in Copper Resources of the World, Vol. 1, p. 31, Sixteenth 
Intern. Geol. Congr., 1935. 



COPPER ORES 7 

western United States and three on the west slope of the Andes in 
South America. These are " disseminated copper deposits," in which 
the copper minerals in the form of small grains are scattered uniformly 
through a large body of rock. The copper minerals in the upper 
portions are in general oxidized, and those lower down are sulfides. 
In the first four deposits to be developed the copper minerals were 
distributed m a porphyry hence the name porphyry coppers. Al- 
though some of these deposits occur in schist or other host rocks the 
name porphyry coppers is generally applied to the entire group. 
Parsons 5 lists the following characteristics of the porphyry copper 
deposits. 

1. The deposit is of such magnitude and shape that it can be mined advan- 
tageously by large-scale methods, either by underground caving or in 
open pits. 

2 The distribution of the copper minerals is so general and uniform that 
" bulk " methods of mining are more profitable than selective methods 
whereby individual veins or thin beds would be stoped separately. 

3. An intrusion of porphyry or closelv related igneous rock has played a 
vital part m the genesis of the ore though the porphyry itself may not con- 
stitute the major part of the deposit The evidence is convincing that 
remarkably large, deep-seated, slow-cooling masses of rock were the source 
of the heat and energy and, directly or indirectly, of the metals in the deposits 
of the present day. 

4. The process known as " secondary enrichment " has operated to con- 
centrate the copper. At New Cornelia the zone of secondary enrichment 
is almost negligible but it exists 

5. The extent of the ore body is usually determined by economic limits 
rather than by geologic structure This is because the copper content 
gradually diminishes as progress is made either downward or laterally from 
the core of an enriched mass At some point which necessarily varies 
with the cost of production at the particular mine, with the price of copper, 
and with other economic conditions a " cut-off " must be made between 
"ore" and "waste." This may be 0.5 per cent copper or it may be 1.5 
per cent in different mines; and, considered literally, it would vary widely 
with respect to the same mine at different times. 

6. The average copper content of the mass is comparatively low (with 
3 per cent as the maximum) and grinding and mechanical concentration are 
necessary to produce a suitable smelter feed, if the ore is sulphide m 
character. 

Some of the important facts about the porphyry copper deposits 
are summed up in Table 2. At the present time the production from 
the twelve porphyry copper deposits accounts for about one-third of 
the world production of copper. 

5 Parsons, A B., The Porphyry Coppers, Am. Inst. Min. and Met. Eng. (Rocky 
Mountain Fund), 1933. 



FROM ORE TO CONCENTRATE 



o 
O 



tt 



*- S *^j) 























B S I 


g 





1 





8 


8 


8 


* 


3 


5 


S 


5 


r-T 


5 


S 


|>I 








eo 








||| 


- 


i-H 


* 




rH 


- 


H 


o v ^- 
















jg 


I 


I 


1 


1. 


I 


I 





S fe ^; 


g 


g 


g 


g 


g 


g 


g 


g cc 

1* g 


| 


8 


" 


8 

CO 


S 


s 


h3 


G ft, 

!" s 


1 


8 


1 


8 


g 





1 




o 














3 M *0 


00 


* 





q 


S? 


S 


S3 


IQ | 
















a 


d S 


G a 


d 9 


al 


t 
C C 


d S 


to 

c d 


o ^ 


o . 




9 2 


O , 


o - 


C ^3 





Metallurg] 
Treatmei 


11 
I 1 


Concentrati 
and Smel 


Concentratl 
and Smel 


2 i 

"i ^ 

S c 

C o3 


Concentrat 
and Smel 


Concentrat: 
and Smel 


"c ^ 

1 = 


| 




| 


s4. 


1 M 


i 


1 M 




4) 
C 

c 


3 

i 


nderground 
block-cavm, 
pen cut c 


pen cut, al 
underground 
block cav in. 


C > 

3 83 

" ^ 


nderground 
block-cavm 


nderground 
block-cavm 


3 
O 

1 




o 


D 


C 


^ 


_j 


P 





c 







3 


S 2 








J -o 


c 


? 


? 


c c 


^ 








2 o 


i II 5 


| 


i 


a w v J 


| 





-^ 2 


g 3 N 

S 


? 

J3 .fl O O 

o u o a 


1 


O 

II 


Ills 

^3 x: o a> 
O O ffl H 


-3 
U 


"3 
xi 



o S 3 


C 














21 
















D, 08 


"ale 














651 


||5 














I HI 


^ 














* 3 .2 -fl 

55 S < 


c 
















- o 
















.p ^a 


\f* 


i^ 


00 


o 


o 


^, 


o 


583 


g 


g 


g 






S 


S 


eS rT -d 
















Q 
















P- 






























8 


CJ 














. H 


o 


^ 


s 


4 




OS 





-2 4< 


1 


il 


5 

*J 

o < 


TS 
0) 


S o 


c 
- o 

jj! 


C 

g 

i* 


^ ^ 


Property 


1 


1 


| 


i 


, 




, 




D 


S 


55 


PQ 


S 


S 


O 



TYPES OF COPPER ORES 



**-! S ^y 

















g 





" 


g" 


g 


8 


Q. $ 
















o 








^ 




<N 








e^ -E 












ill 


CO 


'N 


- 


- 


* 


o ~ 












opS 


| 


| 


| 


| 


| 


II o 


s 


| 


1 


co" 
CO 





H Q 




- 








c 2 o 












| ~. 8 


s 


oo 

8" 


c> 


c? 


"^1 










QO 




3 ^ "5 


CO 




'o 


C* 


^ 


II 




* H 








_ 


to 




W) 


M 


6C 


H C 


C C! 

o 5 




c c 

O 


||| 


c a 

O ^3 


3 1 


M 2 I 


w 


* ^ ^ 

tiS S 




M 2 s 


3i H 


S G M 


c 


S c ^ 

8 S-g 

Jl 


S 12 o, 
5 os g 

o a 


c c & 

1 1 

2 * 

hJ 


TJ 

O 


I *. 






1 


I g, 


1 

c 


^ 1 
S 

II 


3 


3 
O 

c 


1 1 a 1 

Hg g. | 


11 

o> o 


Z 


'c 3 


i 


& 


5 


"3 


" 


D 


o 


O 





D 








0) 


o> 


<u 






o> 








J3 ** <u 


2 




x 


c 


o> C 


2 <fl 





a 


O. 03 

o 5 


^ 0> 


C a 

80 


s M 


O 


O J3 ^C 

See 

O ffl t] 


^2 a 

J 

Q 


3 1 - g 


_4 

6 6 


c 

il s 


5 


3 


-2 




s 


8 "2 
J3 N 


lljl 

o o S <j 


1 g ^ J 

1116 


Si 




11 


c 












*O *a -5 




to 


^ 


CO 


^ 


3 S 


2 


S 


S 


2 


Oi 


0, 












d 












S 
I 


|| 


ef 

3 1 -S 

2* 1 3 


I 

N 

. C 




N 

J c 


II 


-! 




2 o 


o -tj 


<f* 






^? 





? 


ffl 


Cu 


f 


O 


08 


"c 






1 


1 


il 

o* 


(5 

1 
^ 


fe 4$ 

1* 


d 



10 FROM ORE TO CONCENTRATE 

The ore deposits classified under \c and 25 are for all practical 
purposes simply the porphyry copper ores. The other principal copper 
ore deposits of the world (except the native copper ores and some pyritic 
ores) do not fit so readily into a classification based on the metallurg- 
ical treatments used in winning the copper from the ore. Since, how- 
ever, the metallurgy of copper is to be our chief concern, it will be best 
for our purposes to retain this simple tabulation rather than to attempt 
to set up an elaborate classification which would place each ore type 
in its proper geological category. We have already seen that in the 
porphyries we may have both oxidized and sulfide ores in a single 
deposit; the same thing is true in other deposits of copper ore. Also, 
we may find high-, medium-, or low-grade ores all in the same deposit. 
Without further preamble let us briefly consider some of the more 
important of the world's deposits of copper ore. 

Second only to the porphyry deposits in importance are the African 
deposits in the Belgian Congo and Rhodesia. These ores are in a 
belt which extends through the province of Katanga in the Belgian 
Congo and into Northern Rhodesia. The ores usually contain the 
copper minerals uniformly disseminated throughout a mass of rock; 
in this respect they resemble the porphyry coppers, but they differ 
from the porphyries in two important respects: (1) the copper ore beds 
are usually sharply delimited by barren wall-rock, and gradational 
contacts or " economic cut-offs " are rare; and (2) the ore is of much 
higher grade the grades of ore reserves are from 3 to 7.0 per cent 
copper in some of the mines as compared with 1 to 2.0 per cent for the 
porphyries (Table 2). The ores of the Katanga district (up to the 
present time) have been principally oxidized ores with malachite as 
the principal mineral and minor amounts of azurite, chrysocolla, 
cuprite, and native copper. Some of these ores are sufficiently high 
grade for direct reduction smelting; the lower-grade oxidized ores are 
treated by leaching. The ore minerals in the sulfide ores, which are 
typical of most of the Rhodesian ores, are chalcocite, chalcopyrite, and 
bornite, with very minor amounts of pyrite and covellite. These 
deposits differ from most other copper ores in that they are low in 
iron, and pyrite (FeS 2 ) is present only in very small amounts. 

Butte, Montana, has produced more copper than any other district 
in the United States, although now its yearly production is exceeded 
by Bingham, Utah. The Butte ores occur in well-defined veins, 
and the principal copper ore minerals are chalcocite, bornite, and 
enargite, with minor amounts of chalcopyrite and tetrahedrite. Pyrite 
is abundant. 

The Sudbury area in Ontario is one of the most productive regions 



TYPES OF COPPER ORES 11 

not only in Canada but in the world. There are two principal types of 
ores in this district copper-nickel ores and zmc-copper-lead ores. 
This region supplies 90 per cent of the world's nickel and is one of 
Canada's largest producers of copper; it also supplies all the platinum 
produced in Canada. The copper-nickel ore bodies are either masses 
of pure sulfides or mineralized rock containing 10 to 60 per cent sulfides. 
The principal sulfide mineral is the iron sulfide pyrrhotite (Fe 7 S 8 ) ; the 
copper mineral is chalcopyrite, and most of the nickel is in the form of 
pentlandite ((Fe,Ni)S). Platinum is present as sperrylite (PtAs 2 ). 
The copper-nickel ores are by far the most abundant in this district. 

In northern Manitoba, Canada, there are large mineral deposits 
which contain both copper and zinc. The principal ore minerals are 
chalcopyrite and sphalerite (ZnS) with some gold and silver. Pynte 
is abundant. 

The ore of Kennecott, Alaska, consisted of both sulfides (chalcocite 
and covellite) and oxidized minerals (malachite and azurite) in a 
limestone-dolomite gangue; about half of the copper was in the form 
of sulfide and half in the oxidized state. This district has been a 
large producer of copper, but the ore bodies are now exhausted. 

The ore deposits at Cananea, in the State of Sonora, Mexico, con- 
tain the ore minerals chalcocite, chalcopyrite, bornite, and pyrite; parts 
of the deposit are in a limestone gangue. 

The deposits mentioned above will give some idea of the many types 
of ore from \\hich copper is extracted, many of these deposits produce 
low- or medium-grade ore (" milling ore ") as well as a certain amount 
of high-grade (" direct-smelting ore "). While we have included some 
of the most productive districts in the world, it must not be assumed 
that this is an exhaustive list of all the important deposits of copper 
ore. The principal purpose of this information is to set before us 
some facts about various types of copper ore ; this will serve to point up 
the discussions of metallurgical treatment methods which are to follow. 
Let us conclude this section with a brief glance at the two remaining 
types of copper ores native copper and pyritic ores. 

There is only one native copper ore deposit of economic importance 
in the world, and that is located on the Keweenaw PeninsuL , of north- 
western Michigan. The ores are either amygdaloid or conglomerate, 
containing native copper grains. These copper particles range in 
size from a grain that is just visible to the eye to large nuggets; some 
masses have been found which weighed from 40 to 100 tons. In some 
parts of the deposits silver is found in the form of nuggets of native 
silver which is not alloyed with the copper; some of the copper also 
contains arsenic. This district is the second largest in the United 



12 FROM ORE TO CONCENTRATE 

States in total amount of copper produced; in yearly production, how- 
ever, it is now well below several of the other large producing districts. 

The best known of the pyritic ores are the deposits of Rio Tinto in 
the Province of Huelva, Spain. The ores are massive pyrite containing 
chalcopyrite in the form of minute scattered grains, or threads and 
strings in crevices. Ore containing about 2 per cent copper is mined 
as copper ore, and the lower-grade or copper-free pyrite is mined and 
used in the manufacture of sulfuric acid. The mines at Rio Tinto 
have been exploited since Phoenician and Roman times. In the United 
States there is a deposit at Ducktown, Tennessee, where the ore con- 
tains chalcopyrite disseminated in massive pyrite and pyrrhotite. 
This deposit, like that at Rio Tinto, is exploited both as a copper mine 
and as a sulfur mine (for the manufacture of sulfuric acid). 

Byproducts of Copper Ores. Very often copper is not the only 
commercial product obtained from copper ores; sometimes the by- 
products are of minor importance, but occasionally their importance 
may equal or exceed that of the copper itself. The manner in which 
the copper is separated from its byproducts depends upon the nature 
of the association of the substance in the ore deposit, as we may see 
by a few random examples. The Butte district is a large producer of 
zinc as well as copper, but here it has been possible to mine the two 
ores separately, so that the problem is quite simple. At the Fhn Flon 
mine in northern Manitoba, however, copper and zinc sulfides arc so 
intimately associated that the ore must be ground and the two sulfides 
mechanically separated by ore-dressing processes. Many copper ores 
contain recoverable amounts of silver, and this metal will follow the 
metallic copper through all the stages of milling and smelting and is 
separated from it only by the final refining operation. Some of the 
important byproducts of copper ores are listed below. 

Nickel. As mentioned above, about 90 per cent of the world's nickel 
is produced from the copper-nickel ores of the Sudbury, Ontario, dis- 
trict. The two metals are separated in either the smelting or refining 
operation as it is not possible to make a complete separation of the 
copper and nickel minerals by ore-dressing methods. Some of these 
ores are smelted to yield directly a natural alloy of nickel and copper 
Monel metal. This alloy contains approximately 68 per cent nickel, 
28 per cent copper, and 2 per cent iron. 

Silver. Many copper ores contain silver, and the metal is usually 
found in the form of sulfides associated with the copper sulfides; oc- 
casionally native silver is found. Silver follows the copper through 
all the stages of its metallurgical treatment and remains alloyed with 
it through the fire-refining operation. Electrolytic refining methods 



BYPRODUCTS OF COPPER ORES 13 

are used to separate the silver from copper. In 1936, 6 17,388,289 
ounces of silver was recovered from copper ores mined in the United 
States; this represented 28.46 per cent of the total domestic silver 
production. 

Gold. Many Copper ores contain gold as well as silver; it is prac- 
tically always found as native gold associated with the sulfides; it be- 
haves like silver in the smelting and refining operations and, like silver, 
is separated from the copper in the electrolytic refining operation. 
Most gold-bearing copper ores contain only small amounts of gold, 
but such large tonnages of ore are treated that the gold produced makes 
a respectable showing. In 1936 7 domestic copper ores yielded 379,159 
ounces of gold, or 10 03 per cent of the United States production. 

Platinum. Platinum and associated metals of the platinum group 
(palladium, oFmium, iridium, ruthenium, and rhodium) are found in 
most copper ores which contain the other precious metals (gold and 
silver). Often they are present in minute amounts, as in the Butte 
ores, but they follow the gold and silver through the process and are 
eventually recovered when the gold and silver bullion is parted and 
refined. The copper-nickel deposits of the Sudbury district contain 
notable quantities of platinum and related metals; in 1937 8 the re- 
fineries treating the base metals from these ores produced 139,361 
ounces of platinum and 119,867 ounces of palladium, rhodium, and 
other metals of the platinum group. This represented 44 per cent 
of the world's production of platinum. The copper ores of Katanga 
also yield platinum group metals, and in 1937 9 the copper refinery of 
the Union Mmiere du Haut Katanga reported the production of 12,860 
ounces of palladium and 2570 ounces of platinum. 

Molybdenum. Molybdenite, MoS 2 , is found in small quantities in 
some copper ores, and recently it has become feasible to separate a 
high-grade molybdenite concentrate during the milling operations. 
Three large copper mines have already become important producers 
of molybdenum Utah Copper, at Bingham, Utah; Chino, at Hurley, 
New Mexico; and Greene Cananea, at Cananea, Sonora, Mexico. In 
1937 10 the United States produced 92 per cent of the world's output of 
molybdenum, 29,419,000 pounds out of a total of 32,000,000 pounds. 
Of this amount Utah Copper produced 4,912,569 pounds, and Mexico, 

6 The Mineral Industry During 1937, Vol 46, p. 250, McGraw-Hill Book Co., 
New York. 

7 Idem, p. 250. 

8 Idem, p. 487. 

9 Idem, p 486. 

10 Minerals Yearbook, 1938, p. 563, U. S. Bur. Mines. 



14 FROM ORE TO CONCENTRATE 

the second largest producer of molybdenum in the world, produced 
about 1,200,000 pounds entirely a byproduct of the Cananea cop- 
per ores. 

Cobalt. Cobalt is associated with the copper ores of Katanga, 
Northern Rhodesia, and Sudbury, Ontario. The world's production 
of cobalt in 1937 ll was about 2800 metric tons, of which probably 
60 to 70 per cent was obtained from cobaltiferous copper ores. 

Lead and Zinc. Either lead or zinc or both may be associated with 
copper sulfide ores in the form of galena, PbS, and sphalerite, ZnS 
When these are associated with copper ores, and it is not possible to 
mine them separately, the minerals are mechanically concentrated by 
ore-dressing methods into copper, lead, and zinc concentrates, each of 
which is treated separately. Although a considerable tonnage of 
lead and zinc is produced from copper ores, it is small compared with 
the total production from lead, zinc, and lead-zinc ores 

Arsenic. Practically all of the world's arsenic is obtained as a by 
product from either lead or copper smelters Arsenic-bearing minerals 
are associated with the sulfides, and when the ore or concentrate is 
roasted, the arsenic is carried off in the smoke in the form of volatile 
As 2 3 . This compound is recovered from the smoke, purified, and 
marketed as " white arsenic." 

Sulfuric Acid. The source of much of the S0 2 which is used in 
making sulfuric acid is pynte, FeS 2 . The pyrite is oxidized (burned 
or roasted) to produce the SO 2 , which is then further oxidized to S0 a 
and dissolved in water to form H 2 S0 4 . In some of the large pyritic 
deposits containing only about 1 per cent copper, the ore is usually 
mined primarily for its sulfur content, and the copper is a byproduct. 
Many copper smelters maintain sulfuric-acid plants, using as raw 
material the S0 2 gas obtained by the roasting of all or part of their 
sulfide concentrates. 

Tenor of Copper Ores. By " tenor " or " grade " of a copper ore is 
meant simply the copper content of the ore expressed in per cent. We 
have already said enough about copper ores to indicate that it is 
impossible to state what the lower limit should be in order that a given 
deposit might be classed as a commercial ore. An ore containing 2 per 
cent copper might conceivably be classed as high-grade in one mine 
and as waste in another. Modern methods of treatment in both mining 
and metallurgy have made it possible to treat deposits of lower grade. 
In the period 1851-1860, the average tenor of copper ore mined through- 
out the world was 20 per cent copper; in 1914-1930, the average grade 

11 Minerals Yearbook, 1938, p. 558, U. S Bur Mines. 



CONCENTRATION OF COPPER ORES 15 

of all the world's copper ore was 1.5 per cent. 12 This is a startling 
change to take place in only 70 years. It appears, however, that this 
downward trend may be checked or even reversed because of the 
development of the large and relatively high grade deposits in Africa. 

Gangue Mineral* in Copper Ores. In addition to the valuable 
minerals or " values " found in copper ores there are the worthless or 
" gangue " minerals which accompany them, and, of course, if the 
average copper ore contains only 1 or 2 per cent copper, the bulk of 
the ore as mined must consist of these gangue minerals. Quartz is the 
predominant gangue mineral in many vein deposits; pyrite is abundant 
in most deposits, although m some it is exploited for its sulfur content, 
in which case it is really not a gangue mineral. Limestone or dolomite 
is found as a gangue mineral in a few deposits, and in the porphyry 
deposits the gangue minerals are the various silicate minerals which 
make up the host rock The metallurgical treatment which a given 
ore is to receive is often determined by the nature of the gangue 
minerals ; for example in treating the Kennecott ores, it was necessary 
to employ ammonia leaching rather than sulfuric-acid leaching because 
of the large amount of acid-soluble limestone in the gangue The 
principal gangue constituents of several typical copper ores are shown in 
Table 3. The examples listed in this table illustrate the fact that 
copper ores vary greatly in composition, and as we shall find later, each 
ore requires its own combination of metallurgical operations to win 
from it the copper and associated values at the maximum profit, 

CONCENTRATION OF COPPER ORES 

Most of the copper ore mined today is treated by ore dressing 
processes; low-grade oxidized ores are treated directly by leaching, 
and some sulfide and oxidized ores are sufficiently high grade for 
direct smelting, but the bulk of all copper ore is first dressed to put 
it into shape for more economical extraction of the copper and other 
valuable elements. The great advances in ore dressing have been 
made since 1900, and they have had a profound effect on the metallurgy 
of copper. Ore dressing methods have made possible the exploitation 
of the immense deposits of low-grade sulfide ore; but without the re- 
sources of modern milling, these would be masses of worthless rock and 
not copper ore. Not only has ore dressing had its effect on mining 
methods and copper ore reserves by making it possible to exploit 
lower-grade deposits, but it has had far-reaching effects on the pyro- 

12 Furness, J. W., Development of the Copper Industry, in Copper Resources of 
the World, p 2, Sixteenth Internat. Geol Congr., 1935. 



16 



FROM ORE TO CONCENTRATE 



tf 
w 



c- 

o 
O 





o " ti i "3 


h ^ 




fe? 9 














O 3 e3 w ?S 


s 




w ** 




" 2 oT 3 


a 








5 ^ 1 | 1 


. a 

s& 


| 

o 


111 1 

^1^ o 


1 


*1 ? & 


l 




Q 


3, Q. *S 
'S s a 




Ij H 


!i 


O 

"e8 


-2 ^ a be 





ll s Hi 
ills ail 

2 c c S > o -73 
SB 41 2 8 E 

H co 


High-grade le 
chite pnnci 


O 

03 

1 


rt "fl ^ ^ g 

"o -2 a -3 -< 
E S 




00 


s 






e 










0) 


^ 


o 









? 


s 






8 o 


$ 


O5 

<N 






I 5 


a ** 


^ 






W 










CJ 










SS M 




IN 






! 


r^ as 











K 


!: 


oo 

00 


o 


3 


iO >O 


S 


3 


co 

<N 




00 O5 







i 


VI 


*s o 




- 1 


2 


V 


W 


s 


oc 


>0 
(N 


fa 


-i co 


CO 


<N 


. S 


HI 


w o 

71 O 






"2 + -S S 3 o > 

|52 i g ^ 3 


$ ^ 








g ^. o "^ c 


s l 


ti 3 








s-i- 






CJ 


8 S 


5 2 i- 






O 


CO ^ 


^ <*. 






o 


- ^ 








O 


^ N 




cs S 




CO W 




3 ~ 


_ co 


co 


*$ O 


T44 


O x 


^ 3 "2 


C5 


"d t3 


CO <N 




o R 




o S 






H 




H O 






& 










o 








o 


2 




_ 


v-' 


G 


^ c 


CO 


2- 


S g 


^ 


s ~ / *v 


^~^ 


55 


p- 




G -~ 




a 






r 




JS 


1 1 




i 




3) 
1 


9 


fi 


? 






oj 


J 


S 




S 


J 







<a 


g 


>, ^ 


5 


1 


I 


1 

o 


3 5 




J5 


U< 


55 


CO S 



CONCENTRATION OF COPPER ORES 



17 





5 


2 




^ ? 


" 




i 




03 

C 


CJ 

p 




Remarka 


Mill feed, principal sulfides 
chalcopynte and pynte \* 


3 

bO 
O) 

o 

8 

*rt 
J3 

<y 

s 


a 
- 
2 

1 

N 

05 

3 
O* 


Milling ore, contains 51 


T 

1 

g 

<M 

(N 



O 
M 
t_ 

J; 

a 


68% chalcopjnte 


g 1 

1 

3 
c 
C 

0. 

fi 

o 

bfi 

C 

^s 

a 


mineral brochantlte. 


Amygdaloid ore, containing 


O 

a 

4) 

J3 

S 
S, 

a 



CD 


mineral 

















^ 


o 








0) 














O 











^ 














O 


n 








O 
















^ 

























X 








? 

















































o 

bC 

























C 
03 c 

o o 


o 












O 










^ 


o 












CD 










(5 


CO 






CO 

















O> 









00 






00 
co 












h- 






o 


















<N 






% 




, 














^ 






o 






s 










fe 


^ 






3 

















m - 


o 

CD 


OS 






in 


"$ 3 












2 'a S 


O 







c^ 




c3 "B v 












a; 2 ^ 


o 













0, H "g 












PH ^ 


be 


3 




tf) 


3 


-? OJ 














-Jj 


-I" 1 




*? 


-f 


O 


























rvi 


^ O 








llg 

















8S23 

O O 1-" O 










co 


oo 
o 

o 











- 


^ o 
o ^ 


>0 






O 


























"3 


<u 

T3 




(N 






13 


S 'S 


'- 








E2 













o 


lo 










1 


j^ 













s 


o 












5 




00 






K 


a 


^ 








> 


T3 




g 









o 


Fi 

























-C 








>^ 






W 






O 










Location 


! 






1 

S 

a 

efl 

-d 
a 






uicamata, 


r2 


| 

OQ 


1 

ja 








s 






fi 

^o 






6 





^ 


^ 





2 



.0-088 

CO *-" O5 



18 FROM ORE TO CONCENTRATE 

metallurgy of copper as we shall have occasion to notice. The principal 
reason for the replacement of the blast furnace by the reverberatory 
furnace in copper smelting is the fact that the reverberatory furnace 
is more suitable for treating flotation concentrates. Roasting of copper 
ores ajid concentrates has as its primary function ffie lowering of the 
sulfur content. But in a few copper concentrates it has Tieen possible 
to remove pyrite, the principal source of sulfur, by ore-dressing 
methods, thus lowering the sulfur content to such a degree that roasting 
was unnecessary. 

Ore dressing is a series of processes by means of which the constituent 
minerals in ore are mechanically separated into two or more products. 
No chemical change takes place in any of the constituents of the ore 
if copper is present as chalcopyrite in the ore it remains as chalcopy- 
rite in the concentrate. The mineralogical analysis of an ore, and the 
size and association of the individual grains are the primary factors 
which govern the choice of ore dressing methods. The two fundamental 
processes in all ore dressing operations are: (1) comminution, or 
crushing and grinding, to liberate the individual mineral particles, and 
(2) concentration by means of which the comminuted ore is mechan- 
ically separated into two or more fractions. It is beyond the scope of 
this work to consider in any detail the subject of ore dressing; we shall 
be content to consider only two topics the advantages and limitations 
of ore-dressing methods in the processing of copper ores, and the nature 
of its products, or concentrates, which must be smelted to recover the 
contained metals. 

The simplest type of ore dressing operation is a two-product separa- 
tion in which only one concentrate and one tailing result Suppose, for 
example, that we are treating a native copper ore containing 2 per cent 
copper and that the copper is the only valuable mineral ; the rest of the 
ore is siliceous gangue. A perfect separation would yield a concentrate 
containing all the copper and only copper and a tailing containing all 
the siliceous matter and none of the copper. One ton of ore would 
yield 40 pounds of concentrate assaying 100 per cent copper and 1960 
pounds of barren tailing to be discarded. Such a perfect separation 
is impossible of attainment, but these figures represent the limit which 
might be approached rather closely. To take another simple example, 
let us assume that we again have a 2 per cent copper ore with a 
siliceous gangue but that all the copper is in the form of chalcopyrite 
(CuFeS2) ; again let us assume that we can make a perfect physical 
separation of the copper minerals from the gangue minerals. The 40 
pounds of copper in one ton of ore is contained in 116 pounds of chal- 
copyrite which assays 34 5 per cent copper. Therefore, we should 



CONCENTRATION OF COPPER ORES 



19 



have 116 pounds of concentrate assaying 34.5 per cent copper and 1884 
pounds of barren tailing. This is also a theoretically perfect separa- 
tion, but obviously the concentrate produced in the second case is less 
desirable than that in the first one; this is due to the ore minerals, and 
there is nothing that ore dressing methods can do about it. 

Now let us consider what we might expect to attain practically in 
the dressing of these two simple ores The tabulations given below 
represent results which might reasonably be expected. 



CASE 1. 



TABLE 4 

NATIVE COPPER ORE 





Weight 
(pounds) 


Assay 
(per cent Cu) 


Weight of Contained 
Copper (pounds) 


Heads 
Concentrates 
Tailings 


2000 
54 3 
1946 


2 
70 
103 


2000 X 02 = 40 
54 3 X 70 = 38 
1946 X 00103 - 2 



CASE 2. CHALCOPYRITE ORE 



Heads 


2000 


2 


2000 X 02 = 40 


Concentrates 


131 


29 


131 X 29 = 38 


Tailings 


1869 


107 


1869 X 0.00107 = 2 



Note that the practical results differ from the theoretical values in 
two respects. (1) The concentrate does not have its theoretical 
assay value but is always lower, hence there must be some of the 
gangue mineral in it, and (2) the tailing is not barren but contains 
some copper. In order to measure the effectiveness of a concentrating 
operation it is necessary to consider both of these. The effectiveness of 
the process in recovering the valuable metal is measured by the 
recovery, which is the per cent of the total amount of copper in the 
heads that is recovered in the concentrate. In both cases the recovery 
is the same, namely 38/40, or 95 per cent. The effectiveness of the 
process as a concentrating operation is measured by the ratio of con- 
centration, which is the ratio of the weight of the heads to the weight 
of the concentrate. In case (1) the ratio of concentration is 
2000/54.3 = 36.8, or 36 8 to 1, and in case (2) it is 2000/131 = 15.3, or 
15.3 to 1. The grade of the concentrate, or its copper assay, will be 
directly proportional to the ratio of concentration for any given ore 
provided the recovery remains constant; of course, however, it would be 
possible to get very high values for both grade of concentrate and 
ratio of concentration by removing only the " cream " of the valuable 
fraction, in which case the recovery would be low. 




Fio. 1. Photomicrographs of Some Copper Ores. 

a, Native copper and gangue. 

b, Pyrite chalcopyrite ore with small amounts of galena (PbS) and sphalerite (ZnS). 

c, Boraite sphalerite pyrite ore. 

d, Pyrite chalcopyrite gold ore. 

G t Gangue; Ga, galena; Cu, native copper; Py, pyrite; Cp, chalcopyrite; Bo, bornite; 
Sp, sphalerite; Au, native gold. 

20 





d 

21 



(Courtesy American Cyanamid Company) 



& *KUM UKU 1U 

It is impossible to get a mineralogically pure concentrate or a per- 
fectly barren tailing in any commercial ore dressing process. As a 
general rule, and regardless of what method of concentration is used, 
the recovery and grade of concentrate bear a sort of inverse relation 
to one another; if we strive to obtain a very high grade concentrate 
we ordinarily suffer greater losses in the tailing, and if we aim to get a 
high recovery and low tailing assays we must be content with a lower- 
grade concentrate. The operator should strive to balance grade of con- 
centrate against the tailing loss so that the process gives the maximum 
profit on the ore being treated, all things considered. 

The discussion thus far has been concerned with two very simple ores 
When the ore contains other valuable metals in addition to copper, 
when pyrite or pyrrhotite is present (which are difficult to separate 
from some copper sulfides) , or when the particles of ore minerals are 
so small that it is difficult to liberate them from the gangue by grind- 
ing, the problem becomes more complex. A little later we shall con- 
sider some examples of the milling of copper ores to illustrate the 
various methods employed. 

Comminution. Most copper concentrators employ large gyratory or 
jaw crushers for primary breaking, secondary crushers or crushing rolls 
for finer crushing, and rod mills or ball mills for fine grinding. The 
principal exception is that in the milling of native copper ores, steam 
stamps or other special crushers are used for crushing the ore because 
the presence of large pieces of the tough native copper make it difficult 
to use conventional types of crushers. The fine grinding may be done 
in ball mills after the ore has been stamped and the coarse copper 
removed. 

Concentration. Flotation is the principal method employed for the 
concentration of copper ores, although some gravity devices such as 
jigs and tables are used; often they are used in conjunction with flota- 
tion. Until recently oxidized copper ores have been difficult to con- 
centrate satisfactorily by any method, and these ores are usually leached 
directly without preliminary dressing. Flotation concentrate, as a 
rule, consists of very fine particles, and the roasting and smelting of 
flotation concentrate presents problems which are not involved in the 
smelting of ore or coarse gravity concentrates. 

Examples of Copper Ore Dressing Practice. No two copper ores 
are alike, nor are any two given exactly the same ore-dressing treat- 
ment; the aim in each case is to divide the crude ore into several 
fractions such that the subsequent treatment or discarding of these 
fractions yields the maximum profit. Let us briefly consider a few 
examples of the milling methods used on some typical copper ores. 



EXAMPLES OF .COPPER ORE DRESSING PRACTICE 



23 



Anaconda. The concentrator at Anaconda, Montana, treats the ores 
from Butte; the copper concentrator has eight sections each of which 
can mill 1500 tons of ore in 24 hours. The ore is crushed successively 
in a primary gyratory crusher, Symons cone crushers, and rolls. The 
roll product is deshmed and sand and slime treated separately. Con- 
centration is by flotation; the shine goes directly to a separate flotation 
circuit; the sand is ground in ball mills and treated by flotation in the 
main circuit. Metallurgical results of the existing flowsheet are 
shown in Table 5. 13 

TABLE 5 

METALLURGICAL RESULTS AT THE ANACONDA CONCENTRATOR 



Product 


Weight 
(per cent) 


Assays 
(per cent) 


Copper 
Recovery 
(per cent) 


Cu 


Fe 


Si0 2 


A1 2 O 3 


Ore 
Mill concentrate 
Slime concentrate 

Total concentrate 


100 00 
13 98 
5 08 


5 50 
30 60 
20 00 


24 1 
25 4 


4 1 
10.3 


1 4 
2 6 


100 00 

77.77 
18 47 


19 06 


27 77 


24 5 


5 75 


1 7 


96 24 


Mill tailing 
Slime tailing 

Total tailing 


69 52 
11 42 


24 
35 




... 




3 03 
73 


80 94 


256 








3 76 



The copper sulfides in the Butte ores are so intimately intergrown 
with pyrite that it would require extremely fine grinding for complete 
liberation the low grade of the concentrate is due primarily to 
pynte-copper sulfide middling grains. Morrow and Griswold 14 have 
demonstrated in the laboratory that by using a modified flowsheet 
involving the rcgrindmg of a low-grade pyrite-copper middling it is 
possible to make a 40 per cent copper concentrate from the Butte ores 
with the same recovery as that obtained at present in the mill. 

Ahmeek. 15 The Ahmeek mill treats the native copper ore from 
the Ahmeek mine of the Calumet and Hecla Consolidated Copper 
Company on the shores of Lake Superior in northern Michigan. The 
ore is an amygdaloid rock containing from 26 to 32 pounds of native 

13 Morrow, B. S , and Griswold, G. G , Production of High-Grade Concentrate 
from Butte Ores: Am. Inst Mm & Met. Eng. Trans., Vol. 112, p. 413, 1934. 

14 Morrow, B S , and Griswold, G G., op. cit. 

15 Benedict, C. H , Steam Stamps Hold Their Own at Ahmeek Mill: Eng. and 
Min. Jour , Vol 139, No. 12, 1938 



24 



FROM ORE TO CONCENTRATE 



copper per ton (1.3 to 1.6 per cent). The coarse ore flows over a 
picking table on the way to the steam stamps, and here the large 
masses of copper (" barrel work " or " mill mass ") are recovered by 
hand. The stamps crush the ore down to about %-inch diameter and 
the stamp discharge passes to trommels with % -inch openings. The 
trommel oversize is treated in bull jigs to remove copper; the jig 
tailings pass to crushing rolls in closed circuit with the trommels. Only 
material which passes the % 6 -inch holes in the trommels escapes 
from this circuit except the copper concentrate from the bull jigs. The 
% 6 -inch product is deslimed and the slimes go directly to flotation; 
the sands pass through another series of jigs. The tailings from this 
second set of jigs are ground in a ball mill and then treated in another 
set of flotation cells. Finally, the flotation tailing sands are passed 
over Wilfley tables to recover any copper that was too large to float. 
The copper concentrates from both sets of jigs are referred to as " high- 
grade " concentrate, that from the Wilfley tables and some jig 
products as " low-grade " concentrate. In addition to the tables used 
in scavenging operations on the flotation tailings, a Wilfley table is 
used in closed circuit with the ball mill ahead of the classifier, where 
it performs the same function as a " unit " flotation cell. Metallurgical 
performance is given in Table 6. 

TABLE 6 
METALLURGICAL RESULTS AT AHMEEK MILL 



Product 


Copper Assay 
(per cent Cu) 


Recovery 
(per cent) 


Mill feed 


1.3-1 6 


.... 


Mill mass 


95. 0=b 


4.7 


High-grade (jig) concentrate 


90.0 


50.4 


Low-grade (jig and table) concen- 






trate 


80.0 


26.3 


Flotation concentrate 


60.0 


13.6 


Final tailing 


075 


5 



Total recovery, 95 per cent, average grade of all concentrates, 78 per cent Cu, ratio of con- 
centration, 60 to 1. 

Roan Antelope. 1 6 The concentrator of 'the Roan Antelope Copper 
Mines in Northern Rhodesia treats from 8000 to 10,000 tons of ore 
per 24 hours; the average analysis of the ore is given in Table 3. 
Coarse crushing is done with two Superior McCully gyratory crushers 

16 Littleford, J. W., Concentrating Operations at Roan Antelope Copper Mines : 
Am. Inst. Min & Met. Eng. Trans., Vol. 112, p. 935, 1934. 



EXAMPLES OF COPPER ORE DRESSING PRACTICE 25 

and fine crushing with five Symons cone crushers. Fine grinding is 
performed by ten Marcy ball mills with one Fraser and Chalmers 
ball mill for regrinding. Concentration is entirely by flotation The 
average Roan Antelope concentrate will assay about as follows: Total 
Cu, 58.51 per cent, *ide Cu, 0.59; Si() 2 , 11 52; A1 2 3 , 4.08; Fe, 527; 
S, 17.78; CaO, 0.15, MgO, 57; residue, 4.12 In general, with heads 
assaying from 3.15 to 325 per cent total copper and 0.15 to 0.25 per 
cent oxide copper, the Roan Antelope concentrator will produce a 
concentrate containing 58 per cent copper, representing 87 to 88 per 
cent total recovery and 90 to 91 per cent sulfide copper recovery. The 
ratio of concentration is about 20 to 1 

Much of the oxide copper is lost in the tailings; more of this could 
be recovered by flotation if desired, but this would result in lowering 
the grade of the concentrate. The oxide copper does not present a 
very serious problem, however, because the amount of it in the ore is 
diminishing as mining proceeds downward from the oxidized zone. 
Smelting costs increase in proportion to the amount of alumina in the 
concentrate, hence it is desirable to keep the alumina in the concen- 
trates as low as possible. By making a high-grade concentrate (56 to 
60 per cent Cu) the alumina content of the concentrate can be kept 
down, but this means some sacrifice of recovery. 

Katanga. The ores mined by the Union Miniere du Haut Katanga 
in the Belgian Congo are largely oxidized ores. These ores are con- 
centrated by various combinations of hand picking, gravity concen- 
tration, and flotation. Where the oxidized minerals can be freed in 
relatively large pieces, hand picking and gravity concentration are 
used; flotation is used on finer material Flotation gives good recovery 
on malachite, but very little of the chrysocolla can be recovered by 
flotation. The following data are taken from an article by Bar- 
thelemy. 17 

The largest mine working on oxide ores is the Kambove, and this 
ore is milled at the Panda concentrator. Ore is first crushed and 
sorted by hand on picking belts; the pickers remove high-grade, which 
is dropped directly into cars for shipment, and the ore remaining on the 
belt is fed to the gravity section The gravity section makes con- 
centrates on both jigs and Wilflcy tables, and the gravity tailings pass 
to the flotation section, which makes a flotation concentrate and a 
final tailing Metallurgical results are given in Table 7; note the 
exceptionally high grade of the mill feed and of the final tailing. 

17 Barthelemy, R. E , Katanga Ores Offer a Variety of Treatment Problems : 
Eng and Min Jour , Vol. 135, No 9, p. 401, 1934. 



26 



FROM ORE TO CONCENTRATE 



TABLE 7 
MONTHLY METALLURGICAL RESULTS, PANDA CONCENTRATOR 





Tonnage 


Copper Assay 
(per cent Cu) 


Recovery 
(per cent) 


Mill feed 
Picked concentrate 
Gravity concentrate 


60,315 
121 

7,799 


11 80 
37 40 
35 10 


6 

38 5 


Flotation feed 
Flotation concentrate 
Flotation tailing 


52,395 
10,498 
41,897* 


8 28 
35 33 
1 50 


52 2 

8 7 



Over-all recovery, 91 3 per cent, over-all ratio of concentration, 3 28 to I, recovery of copper in 
flotation feed, 85 5 per cent. 
6 By difference. 

The ore milled at the Panda concentrator contains oxidized copper 
minerals as the principal economic values. At other mines in the 
district, however, the ores are more complex. A copper-zinc ore 
from the Prince Leopold mine contains about 12 5 per cent total copper, 
2.2 per cent oxide copper, 9.8 per cent zinc, and 9 8 ounces of silver. 
Ore from the Ruashi mine contains cobalt in the form of carrolhtc, a 
copper-cobalt sulfide containing 41.28 per cent cobalt and 15 53 per 
cent copper. In this district is also found the Chmkolobwe mine, the 
largest radium mine in the world. 

Copper Cliff. Shortly after the completion of the 8000-ton con- 
centrator of the International Nickel Company at Copper Cliff a 
description was published by W. T. MacDonald. 18 The ore treated 
in this mill contains copper as chalcopynte, nickel as pentlandite, and 
large amounts of pyrrhotite; the mill feed will average about 45 per 
cent sulfides. This ore contains about 4.4 per cent copper, 2 2 per cent 
nickel, and about $4.00 per ton in precious metals (gold, silver, and 
platinum metals) Two important facts govern the milling of this ore: 
(1) in order to make a satisfactory recovery of the valuable metals it 
is necessary to recover all the sulfides, hence the over-all ratio of con- 
centration can never be much greater than 2 to 1 ; and (2) a certain 
amount of selection is possible, but the mineral association is such that 
it is not possible to make clean copper and nickel concentrates. Since 
it is possible to depress pyrrhotite and pentlandite and float chalcopy- 
rite, the first step is to make a copper concentrate containing about 
half of the copper in the ore; this concentrate contains about 25 per 
cent copper and 1.25 per cent nickel. After this concentrate is re- 

1{i MacDonald, W. T., Selective Flotation Mill at Copper Cliff. Eng. and Min. 
Jour., Vol. 130, No. 9, p 465, 1930. 



EXAMPLES OF COPPER ORE DRESSING PRACTICE 27 

moved, all the remaining sulfidcs (pyrrhotite, pentlandite, and the 
remainder of the chalcopyrite) arc removed as a bulk concentrate. 
The pentlandite is very intimately associated with pyrrhotite. This 
milling method cuts the bulk of the ore about in half by rejecting the 
non-sulfide gangue nmcrals and yields (Da copper concentrate con- 
taining only a small part of the nickel, and (2) a bulk copper-nickel- 
iron concentrate. Concentration is entirely by flotation except that 
tables are used as scavengers 0^1 the tailing from the bulk flotation 
circuit. 

Falconbriflgc 1 (J At the mill of the Falconbridge Nickel Mines, Ltd , 
in Ontario, the mill feed is a copper-nickel ore somewhat resembling that 
treated by Copper Cliff Concentration is by flotation and there is 
no selective action all the sulfides arc removed in a bulk concentrate. 
Flotation recover 98 to 99 per cent of the copper and about 94 per 
cent of the nickel with a ratio of concentration of about 4 to 1. Metal- 
lurgical results are given in Table 8. 

TABLE 8 
COMPOSITE ANALYSES, IN PER CENT, OF FLOTATION PRODUCTS AT FALCONBRIDGE 





Cu 


Ni 


Fe 


SiO 2 


S 


Mill heads 


1 12 


85 


17 4 


39 1 


7 6 


Concentrates 


4 43 


3 18 


30 


12 1 


28 2 


Tailings 


03 


005 


10 1 


48 8 


3 



Noranda.' 20 The ore from the Home mine in Quebec is a copper- 
gold ore (Table 3). Two types of ore are produced from the mine, a 
siliceous ore and a heavy sulfide ore containing more than 50 per cent 
pyrrhotite and 20 per cent pyrite. All the siliceous ore is sent directly 
to the smelter because its silica content is needed for fluxing. Whether 
or not the sulfide ore goes to the mill or directly to the smelter depends 
primarily on its gold content if the gold content is high the ore is 
smelted directly, if it is low the ore goes to the mill. The " cut-off " 
which determines whether or not a given lot is milling or direct- 
smelting ore varies from time to time and depends on several factors. 
In general, direct smelting costs more but it gives almost perfect re- 
covery of the gold and copper; milling followed by smelting is cheaper, 

19 Gronningsater, A , Gill, J R , and Mott, R. C., Metallurgy at Falconbridge: 
Eng and Mm Jour , Vol 135, No 5, p 195,1934 

20 McLachlan, C. G., The Development of Concentrating Operations at Noranda: 
Canadian Mm. Jour, Vol. 55, No 4, 1934; Increasing Recovery from Noranda's 
Milling Ore- Am Inst Mm & Met Eng Trans, Vol. 112, p. 570, 1934. 



28 FROM ORE TO CONCENTRATE 

but the losses of gold in the mill tailing are quite large. Of course it is 
necessary that the feed to both smelter and mill be kept reasonably 
constant so that the entire capacity of the equipment can be utilized, 
and Noranda has developed an elaborate milling process to produce 
the maximum yield from the milling ore We shall consider the 
Noranda process in rather more detail than the others, because it 
illustrates nicely many of the factors which must be taken into account 
in the milling of complex ore and it also shows the interdependence of 
milling and smelting operations. 

The mineralogical analysis of the average ore milled at Noranda in 
1933 is given in Table 9. The aim of the milling process is to produce 
from this heavy sulfide ore a single copper-gold concentrate containing 
as much of the valuable metals as possible and to discard a tailing 
containing the bulk of the silica and iron sulfides. 

TABLE 9 

AVERAGE MINERALOGICAL COMPOSITION OF NORANDA 
CONCENTRATING ORE TREATED DURING 1933 

Mineral Content 

Chalcopyrite 6 8 

Pynte 22.1 

Pyrrhotite 51 5 

Magnetite 5 

Silica and silicates 18 3 

Gold, native and as telluride 15 

Silver, probably as argentite . 35 

Au and Ag expressed in ounces per ton, all others in percentage. 

Concentration is entirely by flotation, and it is possible by the use 
of suitable reagents (1) to float chalcopynte, native gold, and gold 
telluride away from pynte and pyrrhotite, and (2) to float pyrite away 
from pyrrhotite. 

Flotation of chalcopyrite while depressing the iron sulfides recovers 
about 93 per cent of the copper but only about 66 per cent of the 
gold. The 34 per cent of the gold remaining in the copper-circuit 
tailings is largely associated with the pyrite and pyrrhotite; very 
little is found in the quartz. The gold in the pyrite is in the form of 
extremely small particles, many of which are not more than 1 or 2 
microns in diameter (1 micron = 0001 mm) ; gold and gold telluride 
found in the grains of pyrrhotite usually occur in the form of larger 
pieces. 

The tailings from the copper circuit pass directly to another flotation 
circuit where the pyrite is floated off. The tailings, containing prin- 



EXAMPLES OF COPPER ORE DRESSING PRACTICE 29 

cipally pyrrhotite and silicates, pass to the pyrrhotite regrind circuit 
where they are given another grind and passed through a third series 
of flotation cell-,; here another copper-gold concentrate is removed and 
the tailings are discarded. The pynte concentrate is sent to a fourth 
circuit the pyrite regrind circuit where it is subjected to intensive 
regrinding and then passed to flotation cells for the removal of another 
copper-gold concentrate. In other words, the first step is to remove 
most of the chalcopyrite and about 66 per cent of the gold; the tailing 
is then split into a pyrite and a pyrrhotite fraction by means of flotation, 
and each of these fractions is reground and refloated to give copper-gold 
concentrate. All three concentrates are combined and sent to the 
smelter. Recoveries made in each circuit are shown in Table 10. 



TABLE 10 
METAL RECOVERIES AND RATIO OF CONCENTRATION 





Recovery 




Circuit 


(per cent) 


Ratio of 
Concentration 




Copper 


Gold 




Main copper circuit 


92 77 


65 68 


6. 34 to 1 


Pyrite regrind circuit 


3 OcS 


6 12 


66 6 to 1 


Pyrrhotite regrind circuit 


1 05 


7 33 


56 3 to 1 


Over-all 


96 90 


79 13 


5 24 to 1 



Note that the additional treatments of the main circuit tailing 
result in additional recoveries of 4 per cent of the copper and more 
than 13 per cent of the gold without too greatly affecting the ratio of 
concentration. Work is being continued in an effort to further de- 
crease tailing losses, and it is also proposed to treat the pyrite tailing 
by cyanidation to extract more of the gold. 

Magna and Arthur. Located near Bingham, Utah, the Magna 
and Arthur mills treat the ore from the porphyry mine in Bingham 
Canyon. They are similar in size and general flow sheet, and both 
are all-flotation plants. Table 11 gives typical metallurgical data for 
one month's operation at Magna. Note the low grade of the mill heads 
and the large daily tonnage. The Arthur and Magna plants represent 
the largest milling operation in the world. In addition to copper re- 
covery as shown in the table, Utah Copper is now recovering substantial 
quantities of molybdenite 21 by selective flotation and is saving 

21 Minerals Yearbook, 1938, p. 563, U. S. Bur. Mines. 



30 



FROM ORE TO CONCENTRATE 



TABLE 11 
METALLURGICAL DATA AT MAGNA CONCENTRATOR FOR APRIL 1930 





Assays 


Total 

Cu(%) 


Non-Sul- 
fideCu(%) 


Fe(%) 


Insol- 
uble(%) 


S(%) 


An 

(oz/ton) 


Ag 

(oz/ton) 


Heads 
Concentrates 
Tailings 


0.9572 
31.4880 
1046 


0.035 
030 


2.65 
26 01 
2 27 


5 29 


35 20 


0.0090 
242 
0025 


0884 
2 56 
0195 



Average tonnage milled per day 2cS,127 tons 

Average concentrate made per day 764 tons 

Ratio of concentration 36.8 to 1 

Recoveries, per cent: 

Total copper 89.37 

Sulfide copper 92 14 

Non-sulfide copper 16 61 

Gold 72 98 

Silver 78 54 

* Martin, H S , Milling Methods and Costs at the Arthur and Magna Concentrators of the Utah 
Copper Co U S Bureau of Mines Inf Circ 6470, Jul> 19.J1 

additional amounts of gold by passing the flotation tailings through 
burlap-lined launders. 22 

SUMMARY 

All the primary or " new " copper produced in the world comes from 
one of the deposits of copper ore The copper ore deposits differ from 
one another in many respects, among the most important of which 
are the following: 

1. Size of the deposit. 

2. Grade of the copper ore. 

3. Mineralogical nature of the ore minerals. 

4. Mineralogical nature of the gangue minerals. 

5. Presence of other valuable substances or byproducts in the ore. 

6. Geographical location of the deposit. 

With the exception of the native copper ores from Lake Superior, 
all these ores contain their copper in the form of chemical compounds, 
and before metallic copper can be produced from these raw materials 
one or more chemical processes must be utilized to break up the 
chemical union between copper and other atoms and to purify the 
crude copper. We shall be concerned with these chemical processes in 

22 Engelmann, E. W., Recovering Gold from Copper Mill Tailing: Mining and 
Metallurgy, Vol. 16, p. 331, August 1935. 



SUMMARY 31 

later chapters. Where " wet " or leaching processes are employed for 
extracting copper the crude ore itself is the raw material, but when 
copper is extracted by " dry " or pyrometallurgical methods (" smelt- 
ing ") the raw material or feed may be either crude ore or concentrate 
or both. In modern practice only a relatively small amount of the 
feed to pyrometallurgical copper plants is crude ore; most of the 
copper-bearing material entering the plant is concentrate made by 
treating the crude ore by ore dressing methods. 

Ore dressing, or the mechanical separation of crude copper ore into 
(1) a tailing to be discarded and (2) one or more valuable concentrates 
for further treatment, is an important step in the treatment of most 
copper ores. Some of the important fad> about the milling of copper 
ores are listed below. 

1. Concentration is usually the cheapest way of getting rid of the 
bulk of valueless minerals in the ores. It would be economically im- 
possible to employ pyrometallurgical methods on such ores as the 
porphyry coppers, but it is both feasible and profitable to make a high- 
grade concentrate from these ores and to subsequently smelt this 
concentrate. 

2. In many cases ore dressing methods give a partial or complete 
separation of copper minerals from other valuable substances, thus 
greatly facilitating subsequent treatment. 

3. Where the mine is some distance from the smelter, the erection of 
a concentrating plant near the mine lowers transportation costs, since 
only the relatively small bulk of high-grade concentrate need be 
shipped. This fact is not always of great importance in milling copper 
ores, because often the mill and smelter are located on the same site 
(e g., Anaconda, Noranda) . 

4. Flotation is by far the most effective concentration method used 
in the treatment of copper ores; it has long been used with notable 
success with sulfide ores, and recently it has been employed with good 
results in the treatment of oxidized copper minerals. The physical 
nature of flotation ^concentrate (principally the fine subdivision of 
the mineral particles) has required certain changes in pyrometallurg- 
ical processes which were originally used on crude ore Flotation has 
also made it possible to separate copper sulfidcs from pyrite and pyrrho- 
tite in some cases, and this, too, has had its effect on pyrometallurg- 
ical treatment. 

5. The various processes classed under the general heading of ore 
dressing are flexible and easily adapted to a wide variety of ores, as 
shown by the examples cited. Very often it happens that a mill can 
be easily adapted to meet changing conditions, whether these are 
changes in the ore itself or changes caused by economic forces. 



CHAPTER II 
THE EXTRACTION OF COPPER FROM ITS ORES 

Copper is extracted from its ores by various modifications of pyro- 
metallurgical and hydrometallurgical processes, the former being used 
mainly with sulfide, native copper, and some high-grade oxidized 
ores, and the latter almost entirely with low-grade oxidized ores. 
Electrolysis is commonly used m conjunction with both types of 
extraction; the impure copper bullion produced by smelting methods 
is usually refined electrolytically, whereas the solution obtained in 
leaching practice may be depleted of its copper by electrolysis, although 
precipitation upon scrap iron is quite common where scrap iron is 
available. Fire refining is likewise standard practice with all methods 
of extraction. It is generally used to bring copper to a higher degree 
of purity before refining electrolytically, and for electrolytically re- 
fined copper which, although chemically in a highly purified form, is 
nevertheless weak and brittle, fire refining is used to impart the neces- 
sary physical properties, such as ductility, malleability, and strength, 
which are required in the final fabrication. 

Figure 1 gives in flow-sheet form a general outline of the methods of 
extracting copper from its ores. Detailed steps are not included; 
neither is the disposition of various byproducts, such as gas, fume, 
flue dust, anode mud, spent electrolyte, and the like. The precipitation 
of metallic copper from solution by sulfur dioxide gas and several 
other practices are purposely omitted as not being general. 
^/ Pyrometallurgical Processes. Approximately 85 to 90 per cent of 
the world's total output of copper is extracted from its ores by pyro- 
metallurgical processes. There are three distinct, although by no 
means equally important, modifications in use today, based upon 
the character of the copper minerals. 

1. Native copper ores, and the product obtained by concentrating 
such ores, are smelted by simple fusion. The gangue material is re- 
moved by adding a suitable flux, which forms with the gangue minerals 
a molten slag. The slag and the copper are readily separated in the 
liquid state, due to the great difference in their respective specific 
gravities. 

2. Sulfide copper ores are smelted almost entirely by what is known 
as " matte smelting." This type of smelting is so widely practiced in 

32 



HYDROMETALLURGICAL PROCESSES 



33 



the metallurgy of copper and so peculiar to that metal that a separate 
section of this chapter is devoted to its theoretical aspects. 

3. Oxidized copper ores, when of a fairly high grade, that is, contain- 
ing from 6 to 20 per cent copper, may be reduced by coke and carbon 
monoxide in a manner analogous to the reduction of iron ores. The 
product is metallic copper, usually high in iron and other impurities. 
From the black color of the reduced metal this type of smelting has 



Copper Ores 



Native 
Mill 



Grari 



Sulfides 
I 



Oxidized 
I 



Low Grade 



Metallic 


I i 




Tailing ti 1 


i. 1 


Copper 
I 


Massive 
Sulfides 


Veins or 
Disseminated 


Fire 

Btfjmng 


Pyrif't Smelting 


Mjll 



High Grade High Grade 

Black Copper 
Smelting 



(with little coke) Tailing 
I 



Concentrate 
Roaster 




Low Grade 
Leach 

Solution Residue 

Precipitation 



Matte 
low grade 

Blast Furnace 


Slag 


Rpverberatory 


Matte 


Slag 


Matte 
Converter 




Slag Converter 
1 


Blister Copper 
Fire Refining 


Slag 





Copper Electrolytic Scrap 

I I 9^ 

Fire Cathode Cen ^ nt 

Refining topper CoppeP 
fire Refmino 



Electrolytic Refining 

Cathode Copper 

Fire Refining 

Casting 

Fio. 1. General Treatment of Copper Ores. 

been called " black copper smelting." It is being supplanted by leach- 
ing, or, where sulfide ores are available for mixing, by matte smelting. 
Hydrometallurgical Processes. Although by far the greater portion 
of the total copper of the world is produced today by smelting, and 
only from 10 to 15 per cent by leaching, nevertheless the hydrometal- 
lurgical processes are improving rapidly in attractiveness, ease of 
operation, and efficiency of extraction. With the ever increasing prac- 
ticability of roasting flotation concentrate, leaching the roasted ma- 
terial, and then precipitating the copper from solution either electro- 
lytically or chemically, it is not at all impossible that leaching processes 
might some day displace entirely the smelting methods used at present. 
The results of laboratory tests, however, are not so readily attainable 
on a commercial scale. 



34 THE EXTRACTION OF COPPER FROM ITS ORES 

The outstanding advantage of hydrometallurgical processes lies in 
the fact that the solvent acts only upon the copper minerals, and leaves 
the larger mass of gangue material unattacked. In smelting methods 
this gangue must be fluxed and melted, involving the expense of a 
fluxing material, fuel, and a costly furnace. With low-grade ores this 
expense is prohibitive, and as flotation methods for low-grade oxidized 
ores have not yet been brought to an economic status, such ores can be 
treated profitably only by leaching. 

A variety of hydrometallurgical methods are in use today, but in 
general the steps have been standardized as follows: The ore is first 
prepared for leaching, either mechanically by crushing or chemically 
by roasting, or both. The roasting may be oxidizing, sulfatizing, or 
chloridizing The solvent is next brought in contact with the prepared 
ore, either by percolation, which is cheaper, or by agitation, which is 
more rapid and more efficient. The separation of the solution and the 
residue is effected by gravity or pressure filtration ; continuous counter- 
current decantation has been found most convenient. The copper is 
precipitated from solution chemically or electrolytically, and the pre- 
cipitate is melted, refined, and cast. 

v THEORY OF MATTE SMELTING 

The ultimate objective of any method of extracting copper from 
its ores is to obtain, as economically and as completely as possible, the 
copper in a highly purified form and to reject the accompanying ma- 
terials in such a form that valuable byproducts, if present, may be 
recovered. In smelting sulfide copper ores, the aim is to separate the 
copper from the iron, sulfur, and gangue materials. The precious 
metals, such as gold, silver, platinum, and palladium remain with the 
copper until the final step of electrolytic refining. The separation is 
accomplished in three distinct operations, as follows: 

1. Roasting removes a portion of the feulfur, as well as some of the 
volatile components like arsenic, antimony, and bismuth, and oxidizes 
some of the iron to FeO. In pyritic smelting, the roasting may be 
accomplished simultaneously with smelting, and in the same furnace. 
As will be shown later ,\the degree of roasting determines the amount 
of sulfur left in the roasted ore, and this in turn determines the grade 
of the matte, that is, the per cent of copper in the matte, which is 
a measure of the ratio of concentration. 

2. Smelting, either in the blast furnace, with or without extraneous 
fuel, or in the reverberatory furnace, performs as its main function the 
removal of the gangue minerals to the slag and the smelting of the 
copper, iron, and sulfur, plus the precious metals, down to a matte. 



COPPER MATTE 35 

The iron oxide formed during roasting and some sulfur are likewise 
removed. 

3 Converting separates the copper of the matte from iron and 
sulfur by oxidizing the sulfur to the gaseous product SO 2 , which is 
emitted through the mouth of the converter; the iron is oxidized to FeO 
and fluxed immediately by free silica, forming a ferrous silicate slag. 
The liquid slag is separated from the liquid copper in the same manner 
that oil is separated from water, and is poured off. The crude copper 
remaining contains the gold, silver, and other precious metals, as well 
as some arsenic, antimony, selenium, tellurium, iron, nickel, lead, zinc, 
bismuth, and other impurities. These are removed later by fire re- 
fining and electrolysis. 

- Fundamentals of Matte Smelting. The smelting of sulfide copper 
ores (or mixtures of sulfide and oxide ores) to a copper matte is based 
upon the strong affinity of copper and sulfur for each other, as com- 
pared with that of other metals and sulfur, and the relatively weak 
affinity of copper and oxygen for each other. In a reducing or neutral 
atmosphere and at smelting temperatures, copper and sulfur form the 
stable cuprous sulfide, Cu 2 S, either by direct combination or by the 
decomposition of higher valence copper sulfides. There is usually more 
sulfur present than required by the copper to form this compound. 
The excess sulfur, under reducing conditions, then unites with iron to 
form a correspondingly stable ferrous sulfide, FeS. The iron present 
in excess of the amount required by the sulfur is usually in the form of 
ferrous oxide, FeO, and combines with silica to go into the slag as ferrous 
silicate. Cuprous sulfide and ferrous sulfide are miscible in all pro- 
portions in the liquid state In the solid state they form a eutectlc. 
The mixture, whether liquid or solid, is known as copper matte. 

Copper Matte. 1 A matte is any sulfide which has been prepared 
artificially by fusion. Under the ordinary conditions of preparation, 
a copper matte consists of a mixture of cuprous sulfide, Cu 2 S, and fer- 
rous sulfide, FeS. In practice these are usually accompanied by the 
sulfides of such other metals as PbS, NiS, Ag 2 S, as well as gold, arsenic, 
antimony, selenium, and tellurium. For the present purpose it is con- 
venient and not at all impractical to consider copper matte as a mix- 
ture of Cu 2 S and FeS in any proportions of the two substances. It 
is likewise convenient and sufficiently accurate here to use the atomic 
weight of copper as 64, iron 56, and sulfur 32. 

Cuprous sulfide, Cu 2 S, consists then of 2 X 64 parts by weight of 
copper and 32 of sulfur, or 4 parts of copper and 1 of sulfur, or 80 per 
cent copper and 20 per cent sulfur. Ferrous sulride, FeS, consists of 

1 See also Chapter IV. 



36 THE EXTRACTION OF COPPER FROM ITS ORES 

56 parts by weight of iron and 32 of sulfur, or 63.6 per cent iron and 
36.4 per cent sulfur. Inasmuch as the copper and iron are both present 
combined with definite proportions of sulfur, it is necessary to know the 
per cent of only one component of the matte in order to calculate the 
other two. A matte containing, for example, 40 per cent copper 
obviously contains % X 40, or 50 per cent Cu 2 S, and 50 per cent 
FeS. Therefore, 0636X50 = 318 per cent iron present; and 
100 - (40 + 31 8) = 28 2 per cent sulfur. The percentage compo- 
sition of the matte can also be calculated if either the per cent iron or 
the per cent sulfur is known. The calculations may be summarized as 
follows : 

Let x represent the per rent of copper in the matte. Then 

Cu = x 
CujjS = -|;r 
FeS = 100 - .r 
Fe = T 7 r (100 - x) 

S = ir + TT( IO - i*) 
As the sum of the three components is equal to 100 per cent of the matte, 

Cu + Fe + S = 100 
or 

x + A (100 - $x) + \x + T 4 T (100 - fc) & 100 

Let y represent the per cent of iron in the matte. Then 

Fe = y 
FeS = V0 

Cu 2 s = 100 - - 

Cu = f (100 - 

s = ?/ 



Again, as the sum of the three components is equal to 100 per cent of the matte, 

Cu 4- Fe + S = 100 
or 

|(ioo - -y-2/) + v + fa + idoo - -y</) s 10 

The sulfur is distributed between the copper and the iron; simultaneous equations 
are therefore necessary. We know that 

Cu + Fe + S = 100 
and 

JCu + fFe - S = 

From these two simultaneous equations it follows that, letting z represent the per cent 
of sulfur in the matte, 

S = z 

Cu = 1778 - 4.892 
Cu 2 S = (177.8 - 4.89*) 

Fe = 3 89z - 77.8 
FeS = -V- (3-892 - 77.8) 



RATIO OF CONCENTRATION 



37 



To check that 



Cu + Fe + S = 100 
177.8 - 4.89s + 3.89z - 77.8 + z s 100 



Where constant reference is made to the theoretical composition of 
copper mattes, a tc,ble such as recommended by Richards 2 may be 
drawn up, giving the per cent of each component and constituent for 



100 
90 
80 
70 
60 

?50 
S. 

40 
30 
20 
10 



Cu 2 S 



.FeS 



Fe 




10 



20 



30 40 50 
Per Cent Copper (grade) 



60 



70 



80 



FIG. 2. The Composition of Copper Mattes. 

varying percentages of copper Such a compilation is shown in Table 1. 
Figure 2 represents the composition of copper mattes in graphical 
form. 

Ratio of Concentration. The smelting of copper ores to a matte 
is just one step, although a very important one, in bringing the copper 
content of the ore or concentrate to a higher degree of purity by re- 
moving a certain portion of the accompanying impurities. In other 
words, the metal is concentrated into a product of smaller bulk con- 
taining a higher percentage of copper. The ratio of the weight of ore 
or concentrate smelted to that of matte produced is known as the 

2 Richards, J. W., Metallurgical Calculations, p. 471, McGraw-Hill Book Co, 
New York, 1918. 



38 



THE EXTRACTION OF COPPER FROM ITS ORES 



TABLE 1 
THEORETICAL COMPOSITION OF COPPER MATTES, IN PER CENT 



Cu 


Fe 


S 


Cu 2 S 


FeS 


Cu 


Fe 


S 


Cu 2 S 


FeS 





63.64 


36 36 


00 


100 00 


41 


31.02 


27 98 


51 25 


48 75 


1 


62 84 


36 16 


1 25 


98 75 


42 


30 23 


27.77 


52 50 


47.50 


2 


62 05 


35 95 


2 50 


97.50 


43 


29 43 


27 .57 


53 75 


40.2,5 


3 


61 25 


35 75 


3 75 


96 25 


44 


28 64 


27 3(3 


5,5 00 


45.00 


4 


60.45 


35.55 


5 00 


95 00 


4,5 


27 84 


27 16 


,56 25 


43 75 


5 


59 66 


35 34 


6 25 


93 75 


46 


27.05 


26 9.5 


57 50 


42 .50 


6 


58.86 


35 14 


7 50 


92.50 


47 


26 2,5 


26 75 


58 75 


41 2,5 


7 


58 07 


34 93 


8 75 


91.2,5 


48 


25 4,5 


26 5,5 


00 00 


40 00 


8 


57 27 


34 73 


10 00 


90 00 


49 


24.66 


26 34 


01 2,5 


38 75 


9 


56 48 


34.52 


11 2,5 


88.75 


50 


23.86 


20 14 


02 50 


37 50 


10 


55 68 


34 32 


12 50 


87 50 


51 


23.07 


2,5 93 


03 7.5 


30 2,5 


11 


54 89 


34 11 


13 75 


86 25 


.52 


22 27 


2.5 73 


0,5 00 


3.5 00 


12 


54 09 


33 91 


15 00 


85 00 


,53 


21 48 


2.5 ,52 


GO 2,5 


33 75 


13 


53 30 


33 70 


16 2,5 


83 75 


54 


20 6S 


2,5 32 


07 50 


32 50 


14 


52 50 


33 50 


17 50 


82 50 


,55 


19 89 


25 11 


68 7,5 


31 2,5 


15 


51 70 


33.30 


18 75 


81 2,5 


56 


19 09 


24 91 


70 (X) 


30 00 


16 


50 91 


33 09 


20 00 


80 00 


,57 


18 30 


24 70 


71 25 


28 75 


17 


50 11 


32 89 


21 25 


78 7.5 


58 


17 50 


24 50 


72 .50 


27 ,50 


18 


49 32 


32 68 


22 50 


77.50 j 


59 


16 70 


24 30 


73 75 


26 2,5 


19 


48 52 


32 48 


23 75 


76.2.5 i 


60 


1.5 91 


24 09 


75 00 


2.5 00 


20 


47.73 


32.27 


25 00 


75.00 ! 


61 


15 11 


23 89 


70 2.5 


23 7,5 


21 


46.93 


32 07 


26 25 


73.75 i 


62 


14 32 


23 08 


77 ,50 


22 .50 


22 


46 14 


31 86 


27 50 


72 50 


63 


13 52 


23 48 


78 75 


21 25 


23 


45 34 


31 66 


28 75 


71.2.5 


64 


12 73 


23 27 


80 00 


20 00 


24 


44.55 


31 45 


30 00 


70.00 


6.5 


11 93 


23 07 


81 25 


18 75 


25 


43 75 


31 25 


31 25 


68.75 | 


66 


11 14 


22 86 


82 ,50 


17.50 


26 


42 95 


31 05 


32 .50 


67 50 


67 


10.34 


22 66 


83 7,5 


16 2.5 


27 


42 16 


30 84 


33 75 


66 25 ! 


68 


9 55 


22 45 


85 00 


15 00 


28 


41 36 


30 64 


35 00 


65.00 : 


69 


8.7,5 


22 25 


86 25 


13 75 


29 


40 57 


30 43 


36 25 


63.75 ! 


70 


7.95 


22 05 


87 50 


12 50 


30 


39 77 


30 23 


37 50 


62 50 


71 


7 16 


21 84 


88 75 


11 2,5 


31 


38 98 


30 02 


38 75 


61 25 


72 


6 36 


21 64 


90 00 


10 00 


32 


38 18 


29 82 


40 00 


60 00 


73 


5 57 


21 43 


91 25 


8.75 


33 


37 39 


29 61 


41 25 


58.75 


74 


4.77 


21 23 


92 50 


7.50 


34 


36 59 


29 41 


42 50 


57.50 | 


75 


3.98 


21 02 


93 75 


6.25 


35 


35 80 


29 20 


43 75 


56 25 


76 


3 18 


20 82 


95 00 


5 00 


36 


35 00 


29 00 


45 00 


55 00 


77 


2.39 


20 61 


96 25 


3 75 


37 


34 20 


28 80 


40 2.5 


53 75 


78 


1 59 


20.41 


97 50 


2.50 


38 


33.41 


28.59 


47. ,50 


52.50 


79 


0.80 


20 20 


98.75 


1.2,5 


39 


32.61 


28 39 


48.75 


51.25 


80 


0.00 


20.00 


100.00 


00 


40 


31.82 


28.18 


50.00 


50.00 













RATIO OF CONCENTRATION 39 

ratio of concentration. Thus if 10 tons of ore containing 4 per cent 
copper are smelted to 1 ton of matte containing 40 per cent copper, the 
ratio of concentration is 10 to 1. This can likewise be expressed as the 
ratio of the copper contents of the matte and the raw material. 

As classified by lMcrs, ;i an ore consists of two portions, a metallic 
portion and an earthy portion. In the former are included the copper, 
gold, and silver minerals, as well as sulfides, arsenides, and antimomdes; 
in the latter quartz, limestone, and oxides of iron and manganese The 
valuable materials for which the ore is being exploited are usually found 
in the metallic portion of the ore, whereas the earthy portion is gen- 
erally worthless. Some of the constituents of the metallic portion, as 
for example barren pynte, may likewise be worthless, but the above 
classification has been made on the basis that most of the metallic 
portion, on smelting, goes into the matte and the earthy portion goes 
into the slag In most deposits of copper ore the earthy part is 
present in a much larger proportion than the metallic part An obvious 
desideratum is to remove the usually larger, valueless earthy portion 
from the smaller metallic portion which contains the valuable materials. 

The art of ore dressing has made such a step quite feasible by 
mechanical means and selective flotation has gone farther and made 
it possible to separate the valuable metallic minerals from the barren 
ones. A flotation concentrate of high copper content can now readily 
be obtained, in some cases attaining almost ariy desired percentage of 
copper up to that contained in the pure ore minerals Until methods 
of treating the practically pure copper minerals or even the highest- 
grade flotation concentrate directly in the converter have been per- 
fected, it usually will be found advantageous to leave a certain amount 
of pyrite with the concentrated ore, in order to furnish fuel for the con- 
verter in the form of ferrous sulfide in the matte. 

Within certain limits, which will be discussed in a later paragraph, 
the greater the ratio of concentration attained in smelting the more 
economical will be the subsequent operation of converting, based upon 
the cost per ton of charge smelted. Thus, assuming that the cost of 
converting copper matte is $4 00 per ton of matte and that this matte 
has been obtained from a smelting operation working on a ratio of 
concentration of 8 to 1, the cost of converting ore or roasted concentrate 
is 50 cents per ton. That is, 8 tons of ore have been smelted to 1 ton 
of matte, which oost* $400 to convert, or 50 cents per ton of ore. If, 
on the other hand, the ratio of concentration has been only 4 to 1, then 
4 tons of ore have been smelted to 1 ton of matte, which, we will again 

3 Potors, E. D , Principles of Copper Smelting, p 5, McGraw-Hill Book Co., New 
York, 1907. 



40 



THE EXTRACTION OF COPPER FROM ITS ORES 



Tailings 
1635 Lbs. 
.25% Cu. 




Concentrating 

2000 Lbs. Ore 

5% Copper 



CHART SHOWING MATERIALS ELIMINATED 

IN THE CONCENTRATION AND SMELTING 

BUTTE MINES ORE AT ANACONDA 



Roasting 
To Gases 



Reverberatory 
Smelting 



62 Lbs. 





Casting 



26% Cu. 31% Cu. 46* Cu. 98 9% Cu. 99.4% Cu. 

(Courtesy Anaconda Copper Mining" Company) 

FIG. 3. 



SMELTING IN A REDUCING ATMOSPHERE 41 

assume, costs $4.00 to convert, which results in a converting cost of 
$1 00 per ton of ore. 

The cost of converting a low-grade matte is usually higher than 
that of converting a high-grade matte, so the low-grade matte, obtained 
from a low ratio of concentration in smelting, would result in a still 
higher cost per ton of charge smelted than in the preceding illustration. 
As an example, assume that the cost of converting 1 ton of a 30 per 
cent copper matte is $6 00, while 1 ton of matte con aining 45 per cent 
copper can be converted for $4.00 If the concentiate to be smelted 
contains 15 per cent copper, the ratio of concentration in smelting to a 
30 per cent matte is 2 to 1, and the cost of converting per ton of con- 
centrate is $600 - 2, or $300 By producing a 45 per cent copper 
matte from this same raw material, the ratio of concentration becomes 
3 to 1, and the co-t of converting this matte is $400 - 3, or $1 34 per 
ton of concentrate. 

The desired ratio of concentration can be obtained, as already men- 
tioned, by regulating the amount of sulfur in the roasted ore or con- 
centrate. In blast-furnace smelting, the amount of air entering (the 
blast) affects the amount of sulfur which is removed by oxidation. 
In any event, the amount of sulfur available for the matte determines 
the ratio of concentration and therefore the grade of the matte. Just 
how this is accomplished is shown by a study of the action of heat 
upon various sulfide minerals 

Smelting in a Reducing Atmosphere. When the &ulfides of copper 
are heated m a reducing or neutral atmosphere, in which no oxidation 
takes place, all the sulfur in combination with the copper in excess 
of that required by the compound CiioS will be expelled by volatiliza- 
tion. The resulting cuprous sulfide is rather stable and will fuse with- 
out decomposing Thus eovelhte, or cupnc sulfide, CuS, on being 
heated to a high temperature in the absence of air, will decompose ac- 
cording to the equation 

20uS->Cu 2 S + S' 

yielding molten cuprous sulfide and volatilized sulfur In practice 
this sulfur is carried away with other gaseou? products of combustion 
until it comes in contact with air, when it is oxidized to S0 2 . 

In a similar manner, when the sulfides of iron (or solid solutions) 
are heated in an atmosphere which is not oxidizing, all the sulfur 
present in excess of that required by the compound FeS will be ex- 
pelled by volatilization, and the resulting stable ferrous sulfide will 
fuse without decomposing. Thus pyrite, on being heated to a high 
temperature in the absence of air, loses one-half of its sulfur according 



42 THE EXTRACTION OF COPPER FROM ITS ORES 

to the equation 

FeS 2 - FeS + S 

The more complex associations of copper, iron, and sulfur break 
down in the same way, resulting in a molten mixture of cuprous and 
ferrous sulfides, and eliminating superfluous sulfur. As chalcopyrite 
is an important ore mineral of copper, it may be used as an example. 
Employing approximate atomic weights, chalcopyrite, CuFeS 2 , is 
composed of 35 per cent copper, 35 per cent sulfur, and 30 per cent iron. 
On heating it decomposes into a mixture of Cu 2 S and FeS and loses 
thereby one-fourth of its sulfur according to the equation 

2CuFeS, -> CuoS + 2FeS + S 

I I 1 

matte volatilizes 

The resulting matte contains about 38 per cent copper, so that in 
case the pure mineral were smelted under reducing conditions, some 
concentration would have been effected, although very little 

The sulfur volatilized from FeS 2 and other sulfides which contain 
more sulfur than the stable sulfides is sometimes called " free-atom " 
sulfur. In reverberatory smelting the removal of this free-atom 
sulfur often accounts for the bulk of the sulfur volatilized from the 
charge. 

When an ore contains chalcopyrite and a gangue which can be 
fluxed, a greater ratio of concentration will naturally be obtained, al- 
though the grade of the matte will not be greater than 38 per cent 
copper. 

EXAMPLE 1 

An ore contains 40 per cent chalcopynte. On smelting in a reducing atmosphere, 
what would he 

(a) The weight of matte produced from 100 pounds of ore? 
(6) The grade of the matte 9 
(c) The ratio of concentration? 

Take 100 pounds of ore 

= 40 pounds of chalcopyrite 

= 0.35 X 40 = 14 pounds of copper. 

2CuFeS 2 = CuaS -f 2FeS + S 

2 X 1*4 160 2 X *8 

( fl ) ft! X 40 = 36.5 pounds of matte produced, containing 14 pounds of copper 

i4 X 100 
(j _ _ 35^3 pgj. cen CO pp er m the matte. 

36 5 

100 pounds of ore 38 3'/r Cu in the matte _ 2.74 
36 5 pounds of matte 14% Cu in the ore 1 

= the ratio of concentration. 



SMELTING IN A REDUCING ATMOSPHERE 43 

When free pyrite is present in an ore, the ferrous sulfide resulting 
from the decomposition of the pyrite melts with the matte, diluting it 
and lowering its grade. 

EXAMPLE 2 

An ore contains 25 per cent chalcopynte and 25 per cent pyrite. On smelting in 
a reducing atmosphere, what would be 

(a) The weight of matte produced? 

(b) The grade of the matte ? 

(c) The ratio of concentration ? 

Take 100 pounds of oic 

*= 25 pounds of chalcopynte and 25 pounds of pyrite 
= 0.35 X 25 = 8 75 pounds of copper 

2CuFoS 2 - Cu 2 S + 2FeS + S 

2 X 1-S4 . IbO 2 X 88 

If S X 25 = 22.8 pounds of matte fiom the chalcopyrite. 

FeS 2 = FeS -f S 

120 XX 

^ = 1** * P oun ds f ! <0kS t() the matte from the pyrite. 

(a) 228 -f IS 3 = 41.1 pounds of matte produced, containing 875 pounds of 
copper 

(6) - X 100 = 21.3 per cent coppei in the matte 

100 pounds of ore 21 3' , ( 1 u m the miiUe _ 2.43 
41 1 pounds of matte S 75' < ( 1 u m the oie 1 

= ratio of concent lation 

Ordinarily the analyses of ores are reported in percentages of copper, 
iron, and sulfur rather than m percentages ot the minerals present. 
In such cases the calculations arc made on the basis of the copper 
uniting with as much sultur as is required to form the compound Cu 2 S; 
the remaining sulfur then unites \\ith iron to form ferrous sulfide, FeS, 
and the t\\o sulfidcs melt together to form the matte. Usually an 
excess of iron is present, and this, if not already in that form, is con- 
verted into the basic oxide FeO which is fluxed by the acid flux SiOo, 
forming a ferrous silicate slag Where a raw sulfide is smelted, some 
of the sulfur of the pyrite or chalcopyrite will volatilize and will not 
therefore be available for matte formation. The furnace conditions 
with respect to the amount of sulfur volatilized must then be known 
before the grade of the matte can be calculated. Such examples will 
be cited later on. 



44 THE EXTRACTION OF COPPER FROM ITS ORES 

EXAMPLE 3 

An ore contains 25 per cent copper, 30 per cent sulfur, and sufficient iron for both 
the matte and the slag Assuming that all the copper and sulfur go into the matte, 
what would be 

(a) The weight of matte produced? 
(6) The grade of the matte? 
(c) The ratio of concentration 4 '' 

Calculate the same for this ore when roasted to 16 per cent sulfur and 25 per cent 
copper 

Take 100 pounds of ore 

Raw Roasted 

Copper in the ore 25 25 

Sulfui in the ore 30 16 

Sulfur to Cu 2 S | X 25 = 6j \ X 25 = 6} 

Sulfur to FeS 30 - 6^ = 23f 16 - G\ = 9f 

FeS formed ff X 23f - 65 31 3f X 9f = 26 81 

(a) Weight of matte 25 -f 6^ -f 65 31 25 -f 6^- -f- 26 81 

= 96.56 pounds = 58.06 pounds 

25 25 

(6) Grade of matte -7-7^ = 25.9% Cu ; = 43 1% Cu 

9to )O 08 00 

, N f 100 25 9 100 43 1 

(c) Ratio of concentration or or - - 

96 56 25 58 06 25 

- 1>04 _ l ' 72 

i ~ T" 

The preceding example illustrate- nicely how the grade of the malte 
is regulated by the preliminary roasting operation. It is, of course, 
not altogether practical to assume that all the sulfur in the raw sulfide 
ore goes to the matte. Where an oxidized copper ore ib mixed \\ith a 
sulfide ore, some of the volatilized sulfur may serve as a scavenger to 
recover the copper from the oxidized ore or from the slag. 

Effect of Copper Oxides upon the Grade of Matte. Mixing an 
oxidized copper ore \\ith a sulfide ore produce^ a higher-grade matte 
and is the practical equivalent of roasting the sulfide ore. 

EXAMPLE 4 

A sulfide ore contains 40 per cent chalcopynte Assuming that one-fourth of the 
sulfur in the chalcopyrite is driven off and is not available for making matte, how much 
malachite ore containing 15 per cent copper must be mixed with the sulfide ore in 
order to produce a 50 per cent copper matte under strongly reducing conditions, such 
as those prevailing in a blast furnace using a large amount of coke? (Under these 
conditions the oxygen combined with the copper in the oxidized ore is removed by 
means of the carbon in the coke A later example will show what takes place when 
the smelting operation is performed in a neutral atmosphere, such as prevails in the 
reverberatory furnace.) 



SMELTING IN A NEUTRAL ATMOSPHERE 45 

Take 100 pounds of ore 

= 40 pounds of chalcopyrite, 

= 0.35 X 40 = 14 pounds of copper, and 14 pounds of sulfur 

14 - 1 -/- = 10 .5 pounds of sulfur available for making matte 

A 50 per cent copper matte (from Table 1 or Fig 2) contains 26 14 per cent sulfur 

105 
- = 40 1 pounds of a 50 per cent copper matte formed from 10 5 

2614 pounds of sulfur 

50 X 40 1 = 20.05 pounds of copper in the 50 per cent copper matte 
20 05 14 = 6 05 pounds of copper to be supplied by the malachite ore 

(> 05 

= 40 3 pounds of the malachite ore to be added to 100 pounds of 

the sulhde ore in ord'T to produce a 50 per cent copper 
matte 

EXAMPLE 5 

An oxidized ore contains 25 per cent copper How much pyrite must be added to 
produce a 40 per cent copper matte if smelted under reducing conditions? Assume 
that one-half of the sulfur in the pyrite is volatilized 

Take 100 pounds of the oxidized ore 

= 25 pounds of copper, which will produce 

25 

= 62 5 pounds of a 40 per cent matte 

A 40 per cent copper matte (Table 1) contains 28 18 per cent sulfur 

02 5 X 2S1S = 176 pounds of sulfur required 

Pyrite, FeS2, contains 53 33 per cent total sulfur, 01 under the stated conditions of 
smelting, 20 07 per cent sulfur available for the matte 

7-- - = 66 pounds of p\nte to be added to 100 pounds of the oxidized ore in 
order to pioduce a 40 per cent copper matte 

^ Smelting in a Neutral Atmosphere. When an oxidized copper ore 
is mixed ^ith a sulfule ore, or when copper oxides have been formed by 
roasting a sulfide oie, the grade of matte produced \\i\\ vary according 
to \\het her (lie mateiial is smelted under reducing or under neutral 
conditions In a strongly reducing atmosphere, as in coke blast-furnace 
smelting, the oxygen \\i\\\ the oxidized copper compounds combines 
with the carbon of the coke to form carbon monoxide, or with the 
carbon monoxide to form caibon dioxide. In the reverberatory fur- 
nace the atmosphere is generally a neutral one, and under these con- 
ditions the oxygen with the copper in the oxidized ore combines with 
some of the sulfur present, eliminating that element as gaseous sulfur 
dioxide, according to the chemical reactions 

Cu 2 S + 2Cu 2 O -> OCu + S0 2 

and 

Cu 2 S + 2CuO -* 4Cu + SO 2 



46 THE EXTRACTION OF COPPER FROM ITS ORES 

Eliminating the sulfur in this fashion is again the practical equivalent 
of roasting, for less sulfur remains to form matte, less matte is therefore 
produced with the same amount of copper, and as a consequence the 
matte is higher in copper. In actual practice the copper formed by the 
above chemical reactions may remain as metallic copper and can be 
seen as small particles disseminated throughout the matte. Or it may, 
because of its affinity for sulfur, take that element from the iron 
sulfide present, and the iron remaining will be oxidized by some other 
material in the bath and go into the slag. A matte obtained from a 
given roasted ore, or mixture of sulfide and oxidized ores, may therefore 
be expected to be higher in copper if smelted in the reverberatory 
furnace than if smelted under the strongly reducing conditions of the 
coke blast furnace. 

EXAMPLE 6 

A roasted copper ore contains 24 per cent copper and 16 per cent sulfur. Of the 
copper in this roasted ore, 

10 per cent exists as CuSO4 
5 per cent exists as CuO 
10 per cent exists as Cii2O and 
75 per cent exists as Cu2^ 

Determine the grade of matte produced from this ore if smelted 

(a) In the blast furnace 

(6) In the reverberatory furnace. 

(a) The determination of the grade of the matte obtained on smelting m the blast 
furnace, that is, under reducing conditions, is the same as in Example 3. 

Take 100 pounds of the roasted ore 

Copper in the ore 24 pounds 

Sulfur in the ore 16 " 

Sulfur to Cu 2 S I X 21 = 

Sulfur to FeS 16 - 6 - 10 

FeS formed f f X 10 = 27.5 

Matte formed 24 + 6 + 27 5 = 57.5 

24 
Grade of the matte = 41,7 per cent copper, when smelted 

in the blast furnace. 

(b) In the reverberatory furnace, the oxygen with the oxidized copper compounds 
oxidizes some of the sulfur in the cuprous sulfide The reaction between the copper 
sulfate and the cuprous sulfide is as follows. 



CuS() 4 -f CuaS -* 3Cu + 2SO 2 

64 2 X 32 

from which it can be seen that every 64 parts of copper as CuSO4 eliminate 64 parts 
of sulfur as SO2, or one part of copper as sulfate eliminates one part of sulfur as gas. 



OBJECTIONS TO MAXIMUM GRADE OF MATTE 47 

In the reaction with CuO, 

2CuO + Cu 2 S -> 4Cu + 8O 2 

2 X 64 32 

or one part of copper as CuO eliminates J- part of sulfur as SO2. 
In the reaction with Cu 2 O, 

2Cu 2 O 4- Cu 2 S -> 6Cu + S() 2 

4 X 64 32 

or one part of copper as Cu 2 O eliminates ^ part of sulfur as S02. 
Take 100 pounds of the i oasted ore 
= 24 pounds of copper, 
= 10 X 21 2 1 pounds of copper as OuH()4, eliminating 24 pounds of 

sulfur; 

= 0.05 X 24 = 1.2 pounds of copper as CuO, eliminating 3 pound of sulfur; 
= 0.10 X 24 2.4 pounds of coppei as Cu 2 O, eliminating 0.3 pound of sulfur. 

Of the 16 pounds of the sulfur in the 100 pounds of roasted ore, (2.4 + 0.3 -f 0.3) 
= 3 pounds of sulfur is eliminated as SO 2 . The determination of the grade of the' 
matte produced is then made as before. 

pounds 



Copper in the ore 




24 


Sulfur in the ore 




16 


Sulfur eliminated 




3.0 


Sulfui to matte 


16 - 3.0 


= 13.0 


Sulfur to CuoS 


|- X 24 


= 6 


Sulfur to FeS 


130-6 


= 70 


FeS formed 


|TV X 7 


= 19.25 


Matte formed 


24 +6 + 1925 


- 49.25 




24 





Grade of the matte TcTor = 48 ?5 per cent copper, when smelted in 

the reverberatory furnace. 

u -"" Objections to Maximum Grade of Matte. In the paragraph on 
" The Ratio of Concentration," it was .shown that the higher the ratio 
of concentration in smelting, the more economical would be the results 
with respect to the cost of converting the matte per ton of ore or 
roasted concentrate smelted. As the ratio of concentration can be 
regulated by the amount of sulfur left in the ore, and this in turn can 
be regulated by roasting or by mixing with an oxidized ore, the question 
naturally arises, why not make the maximum ratio of concentration in 
smelting and obtain thereby the most economical results? In other 
words, why not make the highest grade matte possible, even to produc- 
ing the pure cuprous sulfide containing about 80 per cent copper? 

There are four important reasons why the highest grade matte is 
not desirable. 

1. In order to obtain a relatively high grade iratte on smelting a 
sulfide copper ore, an excessively complete preliminary roast would 
be necessary. 



48 THE EXTRACTION OF COPPER FROM ITS ORES 

EXAMPLE 7 

To what sulfur content must an ore containing 8 per cent copper be roasted in 
order to obtain on smelting a 70 per cent copper matte? 
Take 100 pounds of ore 

= 8 pounds of copper. 

o 

= 11.4 pounds of 70 per cent copper matte. 

A 70 per cent copper matte contains about 22 per cent sulfur. 
0.22 X 11.4 = 2 5 pounds of sulfur 

The ore must therefore be roasted to approximately 2^ per cent sulfur. An 
efficient type of roaster will roast 125 tons of a concentrated copper sulfide ore from 
30 per cent sulfur down to 7 per cent sulfur for about 35 cents a ton It would prob- 
ably cost four times as much to carry the roasting operation from 7 per cent down to 
2^ per cent sulfur. 

In order to produce a high grade matte, the cost of the necessary 
roasting operation becomes prohibitive. 

2 In smelting a sulfide copper ore to obtain an unusually high 
grade matte, the accompanying slag is likewise unusually high in 
copper. As has already been shown, in order to produce a high- 
grade matte, it is necessary to roa&t the ore to a low sulfur content 
Copper has a strong affinity for sulfur and a relatively weak affinity 
for oxygen, so the presence of sulfur prevents the oxidation of the 
copper. In the oxide form, copper is decidedly basic and is therefore 
acted upon by the silica present to form a copper silicate. Sulfur is 
furthermore a scavenger for copper, and will remove that metal from 
the slag as cuprous sulfide. It is evident that if the ore has been 
roasted to a low sulfur content in order to produce the required higher- 
grade matte, there will be less sulfur available to protect the copper 
from oxidation and to remove the copper from the slag. 

Furthermore, as a higher ratio of concentration is necessary in 
order to produce a higher-grade matte, more tons of raw material will 
be required to produce 1 ton of matte, and consequently more tons 
of slag will be made. As this slag will have a greater copper content 
than one produced simultaneously with a lower-grade matte, the actual 
loss of copper during the smelting operation will be considerably larger. 

EXAMPLE 8 

Assume that a slag produced in smelting an ore to a 40 per cont matte contains 
04 per cent copper, whereas one with a 70 per cent matte contains 0.6 per cent 
copper. If the original ore contains 10 per cent copper, the production of a 40 
per cent matte requires a ratio of concentration of 4 to 1 ; 4 tons of ore are smelted 
to 1 ton of matte, producing at the same time about 3 tons of slag containing 



OBJECTIONS TO MAXIMUM GRADE OF MATTE 49 

0.4 per cent copper, or a total of 24 pounds of copper lost in the slag. For every 
4 tons of ore smelted, 24 pounds of copper are lost, representing a loss of 6 pounds 
per ton of ore. As the ore contained 10 per cent copper, the percentage loss is 
3 per cent. 

Smelting this same ore to a 70 per cent copper matte requires a ratio of concen- 
tration of 7 to 1 ; 7 tons of ore are smelted to 1 ton of matte, producing at the 
same time about 6 tons of slag containing 6 per cent ropper, or a total of 72 
pounds of copper lost in the tlag. This represents a lo^s of about 10 pounds of 
copper per ton of ore (compaie with 6 pounds of copper per ton of ore on smelting 
to a 40 per cent matte), or a peicrntage loss of about 5 per cent. 

. 3. Smelting to a high-grade matte means producing a relatively 
small amount of matte. In copper smelting, matte is the collector of 
the gold, silver, and other precious metals, just as lead is the collector 
in lead smelting and in fire assaying. A small amount of matte may 
mean an insufficient matte rain to collect all the precious metals. A 
loss of these may more than offset any advantage gained by smelting 
to a higher-grade matte. 

4. A high-grade matte is difficult to treat in the converter. From 
Table 1 and Figure 2, it is to be seen that a matte high in copper is 
high in cuprous sulfide and correspondingly low in ferrous sulfide. As 
will be discussed in greater detail later on, the converting process con- 
sists in blowing air through molten matte, oxidizing the ferrous sulfide 
to SO L > and FeO, this Fe() is fixed immediately by free silica, forming 
a ferrous silicate slag. The slag is poured off, and the next stage con- 
sists in oxidizing the cuprous sulfide to SOo and metallic copper. 
During the converting process the ferrous sulfide acts as a fuel, and if 
the matte is too low in that constituent, as it is in very high grade 
mattes, there will not be sufficient fuel present to keep the temperature 
high enough to carry on the reactions. 

The matte produced in copper smelting operations usually ranges 
between 40 and 50 per cent; a 45 per cent copper matte seems to give 
most desirable results. 



CHAPTER III 
ROASTING 

The ultimate objective of roasting sulfide copper ores is, as has 
been indicated in the preceding chapter, to regulate the amount of 
sulfur in the ore or concentrate in order to obtain, on smelting, a matte 
which can be treated in the converter efficiently and economically. 
Roasting, when used, is obviously a preliminary step to smelting, but 
as a separate and distinct operation may be eliminated altogether under 
certain conditions. 

1. In blast-furnace smelting, for example, it is possible to regulate 
the amount of sulfur oxidized by proportioning correctly the coke and 
the blast; from 10 to 90 per cent of the total sulfur may thereby be 
eliminated. In this process the roasting is actually performed in the 
blast furnace, simultaneously with the smelting operation. 

2. Mixing an oxidized copper ore with the sulfide likewise obviates 
the necessity of roasting, as has been shown 4n the preceding chapter.,, 

3. The preparation of the ore by selective flotation concentration 
may yield a product which, upon simply melting in the reverberatory 
furnace and eliminating some sulfur by the decomposition of certain 
sulfides, may result in a matte having the required copper, iron, and 
sulfur contents. 1 

As the modern trend in copper smelting practice is toward the pre- 
liminary preparation of ores and concentrates followed by smelting in 
the reverberatory furnace, roasting is a generally favored operation. 

Definition. Roasting is a pyrometallurgical process and consists 
simply in heating an ore or concentrate in a certain atmosphere to a 
high temperature (but below the melting points of the mineral con- 
stituents) in order to effect a definite chemical change and usually to 
eliminate one or more components by volatilization. As distinguished 
from calcining, which is usually considered to mean the expelling by 
volatilization of some constituent through decomposition, roasting is 
essentially a process of combination. Depending upon whether the 
combination is with oxygen, sulfur trioxide, or chlorine, the various 
types of roasting are classified as oxidizing, sulfatizing, and chloridizing 
roasting. Sulfatizing and chloridizing roasting are important pre- 
liminary steps only in hydrometallurgical processes, and will not be 

50 



THEORETICAL CONSIDERATIONS 51 

considered here. Although, as we have noted, roasting is funda- 
mentally different from calcining, the roasted product is usually called 
calcine or calcines. 

1 Object. The objects of applying an oxidizing roast to sulfide 
copper ores are 

1. To eliminate a portion of the sulfur as S0 2 . 

2. To eliminate by oxidation and sublimation Certain components, 
such as arsenic, antimony, and bismuth, which may prove detrimental 
to the subsequent extracting and refining operations. 

3. To convert a portion of the iron into the oxide form, in which form 
it will combine with the silica in the smelting operation and be removed 
as a constituent of the slug. 

In order to attain these objects in a reasonable length of time it is 
necessary to consider the conditions which favor a rapid and efficient 
roast. | 

/Theoretical Considerations. In order to obtain a rapid and efficient 
roasting of sulfide copper ores, four conditions are theoretically 
essential. These are (1) sufficient surface of the solid material; 
(2) sufficient oxygen in the roaster atmosphere; (3) sufficient stirring 
of the charge; (4) proper temperature. 

1. Sufficient Surface The sulfide particles should be small in 
diameter in order that the oxygen of the air may come in contact with 
them. The smaller the size of a particle the greater is its surface in 
proportion to its volume, and therefore also to its weight; consequently 
the finer the material is divided the greater will be the surface with 
which the oxygen may come in contact. A one-pound lump of coal, 
for example, has a surface area of about l / 4 square foot. When this 
lump has been ground fine enough to pass through a 100-mesh screen, 
the total surface area of all the particles is considerably over 2000 
square feet. The comparative rapidity and efficiency of oxidation of 
the smaller-sized particles is obvious. Experimentally, the relative 
roasting efficiencies of large and small particles may be tested by 
making a screen analysis of the roasted product and then determining 
the sulfur content of the sized fractions. More sulfur is invariably 
found in the larger pieces, indicating less efficient oxidation ?nd less 
complete roasting. 

If the finer the sulfide ore particles the more rapid and thorough is 
the oxidation, the question may arise, why not pulverize the entire 
charge to the finest degree possible? The objections to such procedure 
are, among others, the expense of such pulverizing, the production 
(during roasting and subsequent smelting) of a large amount of flue 
dust, and the tendency of fine sulfide particles to melt together. If the 



52 ROASTING 

material is too fine, it will pack down to a mass which is impervious to 
air. The fine concentrate produced by the flotation process, passing 
easily through a 200-mesh screen and containing particles which are 
even smaller than the hypothetical 800-mesh, may be roasted without 
difficulty. 

The required fineness of the particles depends largely upon the nature 
of the ore. Some pyrite, for example, decrepitates, and the largest 
particles may have a diameter of % 6 to % i nc h and still roast satis- 
factorily. Some ores seem to roast more readily than others and may 
therefore be more coarsely ground and still result in a rapid and 
efficient roast. 

2. Sufficient Oxygen. As the roasting of sulfide copper ores pre- 
liminary to matte smelting is essentially an oxidizing process, sufficient 
oxygen must be present to supply the sulfur, iron, and other chemical 
elements with the required amount. Not only must the required 
amount be present, but concentration of the oxygen must be high 
enough to insure rapid roasting. A candle will burn in air containing 
20 per cent oxygen; decrease the concentration to 18 per cent oxygen 
and the candle will go out. A human being cannot continue to exist 
in an atmosphere containing less than 16 per cent oxygen. The same 
reasoning applies to the oxidation of sulfide particles; for if the S02 in 
the atmosphere surrounding the ore is more than 4 per cent by volume, 
the subsequent sulfur elimination becomes markedly slight.. 

A sufficient amount of oxygen at an adequate concentration is ob- 
tained by allowing a current of air to pass over the roasting ore. This 
current of air not only brings the oxygen in direct contact with the ore 
particles, but it likewise removes the gases resulting from the oxidation. 

Theoretically it is of interest to consider the possibility of having 
too strong an air current. In the first place, there is a limit to the 
amount of sulfide exposed per unit of time to the action of the oxygen, 
and therefore to the amount of heat liberated per unit of time. A large 
excess of air may absorb so much of this heat that the remainder is 
not sufficient to keep the roasting ore above the ignition point, and the 
oxidation of the sulfide minerals, taking place only above a definite 
temperature, will cease as soon as this temperature is no longer main- 
tained. On the other hand, supposing that the ore particles could be 
stirred efficiently enough so that contact with the oxygen of the air was 
possible to any degree of rapidity, then the heat would be generated at 
such a rate that only a relatively small proportion could be removed 
by convection and radiation. The temperature of the roasting ore 
could then rise until the melting point was reached, and the resulting 
coalescence of the molten sulfide particles would eliminate immediately 



53 



that first essential condition to a rapid and efficient roast, namely, suf- 
ficient surface. 

3. Sufficient Stirring. As roasting is largely a surface effect, it is 
necessary, even when the ore has been crushed to a fine enough degree 
to provide sufficient surface, to bring new unaltered surfaces in contact 
with the oxygen of the air. When sulfide minerals are heated without 
access to air, they either decompose, as in the case of pyrite and chal- 
copyrite, or simply fuse, if the temperature is high enough. By present- 
ing new hot surfaces to the oxidizing influence of the air, oxidation 
takes place before the sulfide can melt. The more efficient the stirring 
operation, the greater will be the capacity of the roasting process. 

4. Proper Temperature. Before combustion will take place, the 
ore particles must be heated to a temperature known as the ignition 
temperature. This temperature varies with the character and size 
of the sulfide mineral particles, ranging from 325 C for 1 mm pyrite 
grains to 800 C and more for 2 mm sphalerite grains. Most of the 
copper sulfide minerals require a dull red heat for ignition. The proper 
temperature for roasting is not the same throughout the entire opera- 
tion At the beginning of roasting, when there is a large proportion of 
easily fusible sulfides present, the temperature must not be so high as to 
melt the particles, else that first requirement for a rapid and efficient 
roast, namely, sufficient surface, will no longer be fulfilled If this 
initial relatively low temperature were maintained throughout the entire 
roasting period, sulfates of the metals would be formed. This is in 
agreement with a rule of thermochemistry, which states that " of two 
or more possible chemical reactions, that one is the more likely to 
occur which evolves the greater amount of heat." The heat of forma- 
tion of sulfates is higher than that of the corresponding oxides, so if 
the temperature is low enough, sulfatos will be formed, with the obvious 
decreased elimination of sulfur. Towards the end of the roasting 
period it is therefore necessary to maintain a higher temperature than 
at the beginning, in order to decompose the sulfates which have a 
tendency to form. Such a decomposition or " desulfatization " temper- 
ature is not difficult to attain, being in the neighborhood of 550 C to 
600 C, at which temperature the sulfates begin to break up. As the 
roasting operation is usually carried on in a moving stream of air, 
the resulting S0 3 is removed rapidly, preventing the building up of an 
increased partial pressure, under which condition the desulfatization 
temperature likewise decreases. As most of the easily fusible sulfides 
have been decomposed and oxidized during the earlier stages of roast- 
ing, it is not necessary to hold the temperature to such a low value 
towards the end of the roasting period. 



54 ROASTING 

Self-Roasting Ores. Some ores, when roasting lias once been started 
by bringing them to the ignition temperature, will continue to roast 
without the aid of any extraneous fuel, simply because the heat pro- 
duced by their oxidation is sufficient to maintain the reacting ore 
particles and the air, as well as the various products of combustion, 
at or above the ignition temperature. Such ores' are known as " free- 
burning " or " self-roasting " ores and are obviously more desirable 
and more economical than those which require additional fuel. 

In order to assist self-roasting in an ore, it is necessary to conserve 
as efficiently as possible the heat liberated by the oxidation of the 
sulfide minerals. This heat is removed eventually from the furnace 
in which the roasting operation is taking place in three major ways. 
(1) It is removed through the furnace walls. The loss of heat, how- 
ever, is minimized by making the walls thick and by making the ex- 
posed area of the walls small as compared to the area of the hearth 
upon which the ore particles are spread for roasting (2) The losses 
in the outgoing gases are curtailed by a countercurrent system. Cold 
ore or concentrate is fed into the furnace through the current of hot 
outgoing gases, absorbing from these hot gases a certain amount of 
heat. (3) In the same manner lo&ses m the hot discharged roasted ore 
are decreased by bringing the ore in contact with the cold incoming 
current of air, where it gives up a certain amount of its heat to the air 
going back into the furnace. 

Through efficient design and operation of roasting furnaces, some 
self-roasting ores may be " dead roasted," that is, roasted down to such 
a low sulfur content that rabbling reveals a black or " dead " ore 
surface, instead of the bright red surface of oxidizing sulfides. For 
matte smelting, however, dead roasting is not desirable, as a definite 
amount of sulfur is necessary to form the matte. 

ROASTING METHODS 

Before turning our attention to the types of furnaces used in roasting 
copper ores and concentrates, let us briefly recapitulate some of the 
items which are necessary for satisfactory roasting; all these must be 
considered in the design of a roasting furnace. 

1. Roasting is essentially an oxidation of the copper and iron sulfides; 
the oxidizing agent is oxygen from the air, and provision must be made 
for an adequate supply of air. 

2. No part of the roaster charge ever becomes liquid; both the feed 
and the calcine are in the form of solid particles. As the reaction 
can only proceed when the particle surfaces are exposed to the oxidizing 



HEAP ROASTING 55 

gases, the roasting material must be continually stirred or rabbled in 
order to expose fresh surfaces to the oxygen. 

3. The principal gaseous product of roasting is sulfur dioxide, S0 2 . 
Provision must be made to remove this from the roaster atmosphere 
before its concentration becomes great enough to slow up or reverse 
the oxidation reactions. 

4. The temperature must be maintained high enough to kindle or 
ignite the sulfides and to keep them above the ignition temperature. 
However, the temperature must never become high enough to cause 
fusion of any of the sulfides. Some sulfide ores and concentrates 
liberate enough heat when roasted to maintain the proper temperature 
without the use of extraneous fuel. When such self-roasting ores are 
roasted, the process is known as autogenous roasting. 

5. In roasting solid particles, a current of air sweeps over the 
material, and there is always a certain amount of dust loss as the 
finer particles are carried away by the stream of gas. An effort should 
be made to minimize dust losses in the furnace itself, and provision 
should be made to recover the dust which escapes from the roasting 
furnace. 

Heap Roasting. I The earliest method used for the roasting of copper 
ores was " heap roasting," in which the ore to be roasted was heaped 
up on a suitably built pile of cord wood, the pile kindled, and the ore 
allowed to burn or roast slowly. This was an extremely crude method, 
and we shall briefly consider some of its outstanding disadvantages; 
heap roasting in many respects is the antithesis of good roasting prac- 
tice as we have outlined it in this chapter, and it will be instructive to 
keep these faults in mind to appreciate how many of them have been 
overcome in the design of modern roasting furnaces. 

1. Although heap roasting could employ cheap and unskilled labor 
for the handling of material, the construction of the heap required 
considerable skill and knowledge of the ore, and the roasting heap had 
to be watched carefully all during the roasting operation to prevent 
overheating and fusion of the sulfides or to prevent the fire from going 
out altogether. Large heaps often required 3 or 4 months to roast 
completely J necessitating large roast yards and careful planning of the 
firing of the heaps in order to yield a continuous supply of roasted ore 
for the smelting furnaces. 

2. Even under the best conditions, the product from heap roasting 
was never uniformly roasted, nor did the final product have a uniform 
sulfur content. Parts of the heap would be dead roasted, other parts 
would melt down to matte, and still other parts of the heap would 
contain " green " or unroasted ore. It was not possible to roast an 



56 ROASTING 

entire charge to a definite sulfur content, and this is the principal ob- 
jective in roasting copper ores and concentrates. 

3. Building the heaps and tearing them down again after roasting 
resulted in high labor costs because of the amount of handling required. 

4. Because of the high sulfur content and low roasting temperature, 
heap roasting often resulted in the formation of large amounts of water- 
soluble copper sulfate. In rainy weather this meant heavy losses 
of copper as well as the damaging of any of the iron parts of handling 
equipment exposed to the corrosive copper sulfate solution. 

5. In heap roasting (and, later, in stall and kiln roasting) natural 
draft was used, and there was no way to confine or control the sulfurous 
gases formed by the burning sulfides. Where heap roasting was prac- 
ticed on a large scale, these gases would kill all the vegetation within 
a wide area surrounding the roast yards. This reason alone would be 
sufficient to prevent the return to heap roasting practice in most 
localities, even if there were no other objections to it. 

6. In roasting a heap of ore it was necessary that the ore contain 
enough coarse material to permit free circulation of air and gases 
throughout the mass. The roasting of finely divided flotation con- 
centrate, for example, would be practically impossible with such 
technique. 

The many disadvantages of heap roasting were recognized from 
the beginning, and efforts were made to develop better methods of 
roasting. Stalls and kilns employing natural draft were used first, 
and although these made the handling problem a little simpler and 
resulted in a better control of the sulfurous gases, they were still 
unsatisfactory. What was needed was a continuously operating 
roasting furnace that would operate as efficiently as possible and 
permit close control of the sulfur content of the roasted products. 

Roasting furnaces eventually developed along two main lines. 
(1) In furnaces for hearth roasting the roasting ore was spread over a 
hearth in a shallow layer and exposed to the oxidizing gases; the ore 
on the hearth had to be continually stirred or rabbled to expose 
fresh surfaces. (2) In furnaces for blast roasting the ore was not 
rabbled, but a blast of air was forced through the mass of roasting 
ore. A still more recent development is flash or suspension roasting, 
which is in some respects an outgrowth of hearth roasting. Flash roast- 
ing has not yet been applied commercially to the roasting of copper ores 
and concentrates, and the use of blast roasting in copper metallurgy is 
largely confined to the preparation of charges for the copper blast 
furnace. Hearth roasting in the multiple-hearth furnace is by far 
the most prevalent method for roasting copper ores and concentrates. 



58 ROASTING 

furnace. The furnace shown in this diagram has seven roasting 
hearths and a drier hearth. The hearths are constructed of refractory 
brick and are arched slightly. The external portion of the furnace is a 
brick-lined steel shell fitted with hinged doors and smaller inspection 
doors on each hearth. The rabble arms are attached to the hollow 
central shaft, and as the shaft is turned by the driving mechanism at 
the bottom, the rabble blades set in these arms plow through the 
material on the hearth, turning it over to expose fresh surfaces to the 
oxidizing gases. These rabble blades are set at an angle, and in addi- 
tion to stirring the ore they move it either toward the center of the 
hearth or towards the periphery. The feed enters the roaster proper 
through an annular opening in the center of the drier hearth ; it passes 
over the next hearth to discharge through holes on the periphery ; it is 
discharged near the center of the third hearth, and continues alter- 
nately in this fashion until it is discharged as finished calcine through 
holes on the circumference of the bottom hearth. The moist concen- 
trate is fed onto the drier hearth near the outside, and the rabble arms 
move it across the hearth toward the center; the discharge to the 
second hearth is luted so that the material forms its own seal and pre- 
vents the escape of gas from the interior of the roaster. Roaster gases 
are drawn off through gas outlets (usually two) located just below 
the drier hearth. Air circulates through the central shaft, and cold 
air is circulated through the hollow rabble arms to keep them cool. 
The air required for roasting is admitted through the central shaft, 
and by means of valves the air supply to each hearth can be regulated ; 
also the entering air may be either cold or preheated. The central 
shaft is insulated against heat and gases by a 4-inch wall of special 
radial tile and a layer of insulating material between the steel shell and 
the fire wall. This insulation together with the natural current of air 
through the hollow shaft keeps the temperature low enough so that 
workmen can enter the shaft without shutting down and cooling the 
roaster. Rabble arms are fastened to the central shaft by means of 
special holders and can be locked or released simply by tightening or 
loosening a single nut inside the shaft. 

The design permits constant and accurate control of the material 
in process at all points. Thermocouples can be installed in the rabble 
arms to permit the operator to read the temperature on each hearth. 
Burners (for fuel oil, gas, or pulverized coal) are usually set in the 
side walls of the furnace to be used when the ore or concentrate does not 
contain enough sulfur to be self -roasting. The operator can regulate 
rate of feed, air supply, and temperature in such a way^as to obtain 
the maximum roasting efficiency for the material being treated. 



BLAST ROASTING AND SINTERING 



59 



The Nichols Herreshoff 
Furnace. Figure 2 is a cross- 
section of the Nichols Herre- 
shoff multiple-hearth roast- 
ing furnace. These furnaces 
are made in various sizes; 
the diameters range from 6 
to 21 feet, and they may con- 
tain from 4 to 12 hearths. 
The rabble arms are fastened 
to the rotating central shaft 
which is supported on a step 
bearing and driven by gears, 
as shown on the diagram. 
The central shaft is a ver- 
tical cast-iron column con- 
structed in sections. It con- 
sists of an inner, cylindrical 
part or the " cold air tube " 
and an outer annular part 
or the "hot air compart- 
ment." Cold air is forced 
in through the cold air tube 
and passes from here into the 
hallow rabble arms, thus 
serving to keep them cool. 
The heated air coming from 
the rabble arms enters the 
hot air compartment, and 
from here it may be dis- 
charged to waste at either the top or bottom of the shaft, or admitted 
to the hearths as preheated combustion air. All rabble arms are in 
parallel, and each arm receives its own supply of cooling air directly 
from the cold air tube. 

BLAST ROASTING AND SINTERING 

The Dwight-Lloyd sintering process is today the most wiueiy useu 
method for blast roasting and sintering. As wr have noted, blast 
roasting is a roasting method in which the charge is held stationary and 
the air for roasting the particles is forced or drawn through the inter- 
stices of the bed of ore or concentrate. The principle of the Dwight- 




(CourUsy Pacific Foundry Company, Ltd.) 

FIG. 2. The Nichols Herreshoff Roasting 
Furnace. 



60 ROASTING 

Lloyd process consists in subjecting a thin bed of fine materials to heat 
developed by combustion of fuel within the bed while the individual 
particles are held in a quiescent state. Of course, in roasting sulfidc 
concentrates the sulfides themselves serve as the fuel ; if the material 
to be sintered contains no combustibles, a small amount of coke 
breeze or other fuel is mixed with the charge Sintering refers to a 
physical change in the material undergoing treatment, in which the 
finely divided particles are converted to a cellular porous sinter cake; 
in sintering sulfide concentrates, the process may also give a complete 
or partial roast depending upon how much of the sulfur is oxidized and 
removed. Material such as finely divided oxide iron ores may be 
mixed with coke breeze and sintered, and such an operation would be 
simply sintering and would not be considered roasting in the sense that 
the word is used in non-ferrous practice. There is, then, a definite 
distinction between sintering and roasting; the first term refers to the 
physical process of agglomerating fine particles into coarse pieces, and 
the second term refers to a chemical change brought about by oxidation 
of the charge (usually meaning the elimination of sulfur by the burning 
of sulfides) . 

The Dwight-Lloyd sintering machine (Fig 4) is a structural steel 
framework supporting a closed track around \\hich travels a series 
of small grate-bottom cars or pallets for currying the charge, driving 
mechanism, suction boxes beneath the upper pallet track sect'on con- 
nected to an exhaust fan for drawing air through the bed, a feed hopper, 
and an igniter for starting combustion of the fuel in the charge. The 
charge is fed to the mixer, where it is moistened, mixed, and worked up 
to a fluffy, air-permeable condition; then it passes to the distributing 
device which delivers it evenly across the full width of the pallets be- 
hind the feed hopper. The hopper carries no storage but is a three- 
sided open-backed leveling plate for maintaining a uniform depth of 
bed in the pallets and for laying the charge on the grates with the 
coarser particles on the bottom and the finer ones on top. As the pallet 
moves from under the feed hopper the charge passes under the igniter 
at the front end of the suction box. The igniter may be fired with 
gas, oil, coal, powdered coal, or even wood; its purpose is to project an 
intense flame on a small area of the upper surface of the bed and 
kindle the fuel in the charge. After passing the igniter the charge 
moves across the suction boxes, where sintering takes place, and is 
finally discharged as finished sinter cake. 

Combustion, in the sintering process, proceeds downward through 
the bed in a relatively thin zone, only a small layer of the charge being 
at the maximum (sintering) temperature at any given time, as shown 



BLAST ROASTING AND SINTERING 



61 




^s +. 




(Courtesy Dwiqht and Uoyu Sintering Company, Inc.) 

Fia. 4. Sectional View of a Dwight-Lloyd Machine. 



62 ROASTING 

in Figure 6, a time-temperature curve taken at a point halfway down 
in the bed of charge. Thus, when the charge on any given pallet is 
half finished, the sintering zone will be found halfway down in the bed 
with everything above it finished sinter. In this narrow sintering zone 




(Courtesy Du*ight and Lloyd Sintering Company, Inc.) 

FIG. 5. Discharge End of Dwight-Lloyd Machines Showing Finished Sinter. 

the charge particles are semi-fluid as a result of the rapid combustion 
caused by the air blast, which has been preheated in passing through 
the still warm sintered zone above. The hot combustion products pass 
on down, in turn preheating the charge in the zone beneath. While the 
charge is brought to fusion for an instant in the sintering zone it does 
not have time to become molten, as the action passes quickly beyond 
any individual particle and the cold air blast following chills it before 
it has time to flow together, leaving the sinter in the form of a cellular 
porous cake (Fig. 5). 

The speed of travel of the pallet train is so regulated that the pallet 
leaves the suction box zone at the moment the charge has been com- 
pletely sintered. The pallets do not completely fill the space on the 
tracks, and when a pallet passes around the discharge curve track 
section it separates from the rest of the line, bumps against the pallet 
aheadj and jars loose the charge, now transformed into sinter, to dis- 
charge the cake into a bin or directly onto a railroad car. The empty 



BLAST ROASTING AND SINTERING 



63 




pallet travels back along the lower track section towards the drive 
sprocket to complete the cycle. 

The Dwight-Lloyd process produces a sinter which makes ideal feed 
for the blast furnace; it is free from fines and dust, strong enough to 
support the weight of the charge, porous, and 
hence readily permeable by the gases; and the 
material is prefused so that it smelts more 
easily. Blast roasting and sintering are not 
used in the metallurgy of copper nearly as 
much as is the multiple-hearth roasting process. 
Most copper smelting is done in reverberatory 
furnaces, and the feed for these furnaces is 
usually either calcine from multiple-hearth 
roasting furnaces or raw concentrate. When 
finely divided material is to be smelted in a 
copper blast furnace, however, the Dwight- 
Lloyd process is used; in this ca^e the principal 
purpose of the treatment is the sintering of the 
charge, and the roasting is merely incidental. 

At the copper-nickel smelter 1 oi Falcon- 
bridge Nickel Mines, Ltd , near Sudbury, 
Ontario, the ore and concentrate are smelted 
to a copper-nickel matte in a bla>t furnace. Fic. 6. Time-Tempera- 
Fine ore and flotation concentrate are sintered ture Curve at a Point near 

4 4, i i ^o i O<M i TA i tr T i i the Center of the Bed on a 

on two standard 42 by 2M inch Dwight-Lloyd T , 1A T1 , ,, ,. 

ri ,. * , J Dwight-Lloycl Machine, 

machines. 1 lie ore contains considerable quan- 
tities of iron suliides in addition to the copper and nickel sulfides, 
and the coarse ore suitable for direct smelting contains about 55 per 
cent sulfides. The feed to the D\Mght-Lloyd machines consists of 
u fines " or ore les.s than % inch in diameter containing about 40 
per cent sulfides, bulk flotation concentrate obtained by milling the 
leaner ore, and flue dust. The charge consists of 66 per cent fines, 
27 per cent flotation concentrate, and 7 per cent flue dust. The 
Dwight-Lloyd charge carries about 19.5 per cent sulfur, and this is 
reduced to 105 per cent in the sinter. Each machine requires about 
14,000 cubic feet of air per minute, uses 0.4 gallon of oil per ton of 
sinter in the igniter, and produces on an average 4.35 tons of sinter 
per hour. Since sulfur forms part of the fuel in the blast furnace, 
the sulfur loss on sintering is considered a drawback rather than an 
advantage; this sacrifice must be made, however, to provide the fuel 
necessary for the sintering action. 

1 Gronnmgsator, A M Gill, J. R., and Mott, R. C., Metallurgy at Falconbridge: 
Emr mid Min Jour Vol. 135. No. 5. May 1931. 



Time-Mmutes 

(Courtfiy D\nght and Lloyd 
Wintering Company, Inc.) 



64 ROASTING 

Another example of the use of Dwight-Lloyd machines is the practice 
employed at the Comston smelter of the International Nickel Com- 
pany. 2 Here the process is also used to sinter fine ore previous to 
blast furnace smelting; resulting sinter plus the coarse ore make up the 
blast furnace charge. There are six 42 by 396 inch Dwight-Lloyd 
machines in this installation, and the capacity of each machine is about 
250 tons per day. Fuel oil is used to ignite the charge, and the sulfur 
content is reduced from about 15 per cent to 10 per cent. 

FLASH ROASTING 

The process of flash or suspension roasting has not had any com- 
mercial application to the roasting of copper concentrates as yet. 
However, it is possible that it may be used in the future, and as we 
shall have occasion to refer to flash roasting in the discussion of certain 
aspects of copper smelting in the next chapter, we shall present here a 
brief discussion of the principles involved. The description and dia- 
gram used for an illustrative example refer to the suspension roasting 
of zinc sulfide concentrate as practiced at the zinc plant of the Con- 
solidated Mining and Smelting Company of Canada, Ltd., Trail, B. C. 3 

In the ordinary hearth roaster, a large part of the roasting takes 
place while the particles are dropping from one hearth to the next; 
each particle is completely surrounded by the oxidizing gases in the 
furnace atmosphere, and combustion is much more rapid thrvn if the 
particles were lying on the hearth and exposed to the furnace gases only 
when turned up by the action of the rabble blades. In the process of 
suspension roasting, all the roasting is done while the sulfide particles 
are falling through a stream of oxidizing gas; obviously the method 
can be used only for roasting finely divided material. 

The flash-roasting equipment used at Trail consists essentially of a 
standard 25-foot diameter Wedge roaster with the central hearths 
removed to form the combustion chamber; a diagram of the equipment 
is shown in Figure 7. 

The wet concentrate feed enters through the hopper (14) and passes 
over the drying hearths (3 and 4). The dry material passes through 
a chute and feeder (17) into a ball mill (16) which is used primarily 
to break up agglomerations formed in drying. An elevator (18) con- 
veys the dried and disintegrated concentrate to the hopper and feeding 
device (15). By means of a combustion air fan (9) and burner (10) 

2 Canadian Min Jour., Vol 58, No. 11, p 683, 1937. 

3 Stimmel, B A , Hannay, W. H., and McBean, K. D., The Electrolytic Zinc 
Plant of the Consolidated Mining and Smelting Company of Canada, Ltd.: Am. 
Inst. Mm & Met. Eng. Trans , Vol. 121. r> 540. 1936. 



CHEMISTRY OF ROASTING 



65 



the powdered concentrate is sprayed into the combustion chamber (2) 
where the burning takes place; the combustion maintains the chamber 
at a temperature of 1650 to 1750 F. The calcine settles out on the 
collecting hearths (5 and 6) , and after being rabbled across these is 




21 



(Mimrnfl, et al , Am In*t Mm it .Vtf Ena Trans., Vol. 181, p. 542, 1936) 

FIG. 7. Arrangement of Apparatus in Trail Suspension Roasting Process. 

discharged as finished calcine (19). If desired the calcine can be 
diverted to the chamber (7) below hearth (6), and by rabbling in an 
atmosphere containing large amounts oi S() 2 a certain amount of the 
zinc oxide can be sulfated. The air used for combustion may be pre- 
heated if neci'bbury. Eight of tliete feiispension roasters at Trail have 
replaced 25 standard Wedge furnaces, indicating a three-fold increase 
in capacity. 

'^CHEMISTRY OF ROASTING 

The roasting operation is primarily one of oxidation of solid material 
by means of oxidizing gases, and .some typical chemical reactions are 
presented below. Pynte and chalcopynte tend to decompose en simple 
heating to yield elemental sulfur and the stable sulfides Cu 2 S and FeS. 



FeS 2 - FeS + S 



2FcS + S 



(1) 

(2) 



Of course any liberated sulfur would be immediately oxidized to S0 2 . 
However, this decomposition breaks up the solid particles and allows 



66 ROASTING 

the oxidizing gases to penetrate the interior. For this reason pyrite, 
for example, usually roasts more readily than pyrrhotite (Fe 7 S 8 ) which 
does not decrepitate with the expulsion of elemental sulfur, and hence 
oxidizes only on the surface. The oxidation of these iron sulfides is of 
importance because a certain amount of pyrite or pyrrhotite is present 
in all copper concentrates which are to be roasted. In general it is the 
burning of the iron sulfides which accounts for most of the sulfur 
elimination the copper sulfides tend to remain as such in the calcine. 
The reactions for the complete oxidation of pyrite and chalcopyrite are: 

4FeS 2 + 110 2 -* 2Fe 2 3 + 8S0 2 (3) 

4CuFeS 2 + 130 2 -> 40uO + 2Fe 2 O 3 + 8S0 2 (4) 

The hot Fe 2 3 and Si0 2 found in roaster products act as catalyzers 
to promote the further oxidation of sulfur dioxide: 

2SO 2 + 2 ^ 2S0 3 (5) 

and the resultant S0 3 gas reacts with metallic oxides to form sulfates 
or basic sulfates according to reactions such as these: 

CuO + SO 3 ^ CuS() 4 (6) 

FeO + S0 3 ^ FeS0 4 (7) 

Fe 2 3 + 3S0 3 ^ Fe 2 (S0 4 ) 3 (8) 

2( 1 u() + S0 3 ^ Cu()-CuSO 4 (9) 

Sulfur must be in the form of SO 2 or S0 3 in order to be eliminated; 
sulfur contained in sulfates or basic sulfates will remain in the calcine. 
All reactions in which S0 2 is formed will slow down as the S0 2 content 
of the atmosphere increases, arid if the roaster atmosphere contains 
more than about 9 per cent S0 2 by volume the roasting practically 
stops. The equilibrium shown in Equation 5 indicates that the SO 3 
content of the gases will increase as the S0 2 content increases; higher 
S0 3 concentrations mean the formation of larger amounts of sulfates 
(Equations 6 to 9) . Higher temperatures, however, drive the reactions 
in Equations 6 to 9 to the left and decompose the sulfates. 

According to recent research on the reactions and mechanics of 
roasting, it appears that sulfides do not oxidize directly to oxides and 
S0 2 , but that the primary reactions result in the formation of sulfates; 
oxides and S0 2 are products of secondary reactions. The following 
summary is taken from an abstract of a paper presented by Mr. Ash- 



CHEMISTRY OF ROASTING 67 

croft 4 before the Sixty -Third Meeting of the American Electrochemical 
Society. 

(a) Reactions in roasting proceed primarily and definitely to the forma- 
tion of sulphates, not oxides, the latter as well as the sulphur dioxide 
evolved, being secondary products, formed by decomposition of the sul- 
phates. Iron oxide acts as an efficient catalyst in the formation of these 
sulphates. 

(b) Iron is probably the only element, or at least the only principal 
element, which, by reason of the great heat of formation of the higher 
oxide, FoyO,, is completely converted to oxide in a sufficiently oxidizing 
atmosphere, yielding the undecomposed acid radical to a basic material 
such as copper oxide or the basic constituents of the gi-ague. 

(c) As a corollary to the preceding statement, iron sulphate is not, as gen- 
erally assumed, per se decomposed normally at a temperature so far below 
the decomposition temperature of copper sulphate that a mere roasting at a 
carefully regulated temperature may be employed to assure complete con- 
version into soluble copper and insoluble iron. 

(d) Formation of cupnc fernte, CuOFe 3 , takes place at any tempera- 
ture above 550 C, when the oxides of copper and iron are brought in 
juxtaposition; at 700 C, such formation is prohibitively rapid. 

(e) The ordinary rabbled furnace is inimical to complete conversion into 
copper sulphate, on account of the following reaction : 

CuS0 4 + CuS + 2 -> 2CuO + 2S0 2 (10) 

In well-rabbled charges the reaction begins at a temperature almost as low 
as that at which the oxidation of sulphur to sulphates starts. It certainly 
proceeds rapidly at 400 C. 

It appears from Mr. Ashcroft's conclusions that the reactions in 
roasting involve the primary formation of sulfates followed by the 
decomposition of these sulfates to yield oxides and gaseous oxides of 
sulfur either because of interaction between sulfates and sulfides as 
indicated by Equation 10 or by the simple decomposition of these 
sulfates on heating (Equations 6 to 9). The rabbling of the charge 
promotes Reaction 10 by bringing sulfide and sulfate particles in con- 
tact. The decomposition or formation of sulfates depends upon the 
dissociation tension of the sulfate in question; this is a pressure (given 
in mm of mercury) which measures the tendency of the compound to 
dissociate; it increases with the temperature. To illustrate the 
meaning of this let us consider an example. The dissociation tension 
of Fc 2 (S0 4 )a at 553 C, is 23 mm; this means that at 553 C the 
reaction 

Fe 2 (S0 4 ) 3 ^ Fe 2 3 + 3SO 3 (11) 

4 Sulphatizing Roasting: Eng. and Min. Jour., Vol. 134, No. 10, p. 420, 1933. 



68 



ROASTING 



is at equilibrium if the partial pressure of S0 8 in the atmosphere is 
23 mm; if the partial pressure of S0 3 is less than 23 mm the reaction 
will go to the right and more of the sulfate will decompose; if the 
partial pressure is greater than 23 mm, the reaction will go to the left 
and more of the oxide will be sulfated. Table 1 lists the dissociation 
tensions of copper and ferric sulfates at various temperatures. 

The temperature and partial pressure of SO-, will determine whether 
or not a given sulfate will decompose, provided that no undccomposed 
sulfides are left in the roaster charge, if sulfides arc present, and the 
charge is being rabbled, reactions such as 10 will take place and the 
sulfates will be decomposed even though the temperature may be too 
low for normal decomposition of the sulfate In this connection, 
Ashcroft's work 5 shows that if a sulfatizmg roast is desired, no 
movement (rabbling) of the charge must take place as long as unox- 
idized sulfides are present. 

TABLE 1 
DISSOCIATION TENSIONS FOR SULFATES 



Fe a (S0 4 )s ^ 


2CuS0 4 ^ 


2CuO SO 3 ^ 


Fe 2 3 -f 3SO 2 


2Cu() SO.? + $O 3 


2CuO + SO 3 


Temp C 


mm 


Temp C 


mm 


Temp C 


mm 


553 


23 


546 


43 


600 


62 


570 


33 


588 


55 


653 


98 


592 


45 


615 


70 


686 


123 


614 


70 


642 


98 


705 


139.^ 


634 


113 


665 


130 


728 


173 


650 


149 


700 


233 


745 


209 


660 


182 


714 


324 


775 


298 


680 


286 


725 


460 


805 


542 


690 


401 


731 


647 






699 


560 










707 


715 











a Liddell, D M , Handbook of Non-Ferrous Metallurgy, Vol 1, p 341, McGraw-Hill Book Co , 
New York, 1926. 

Roasting of copper concentrates in the multiple-hearth furnace may 
be characterized by the following items: 

1. Although sulfates appear to be the primary products formed, the 
high temperatures in the furnace and the interaction of sulfates and 
sulfides cause a decomposition which results in the formation of oxides 
of the metals and gaseous oxides of sulfur. 

5 Sulphatizmg Roasting, op cit , p. 420. 



CHEMISTRY OF ROASTING 69 

2. The bulk of the sulfur elimination is due to the oxidation of iron 
sulfides; the copper tends to remain in the sulfide form. 

Practically all the concentrate treated in multiple-hearth roasters is 
subsequently smelted to matte in reverberatory furnaces, and, as we 
have seen from the examples in Chapter II, the principal factor is 
the ratio of copper to sulfur, as it is this that determines both the 
amount and grade of the matte. All the charge is melted down in the 
smelting furnace, and the compounds of copper, sulfur, and iron react 
one with another to form the Cu 2 S and FeS which make up the matte. 
If copper sulfate is present (Example 6, Chapter II), the interaction 
of sulfatcs and sulfides cause the elimination of some sulfur as SO 2 , 
but with this exception, the actual distribution of sulfur, copper, and 
iron in the compounds which make up the calcine is not of great im- 
portance as far as the smelting operation is concerned. We have also 
noted that the amount of sulfate present in the ordinary calcine is quite 
small. 

In recent years, however, there has been considerable interest in the 
problem of developing a technique for roasting and then leaching copper 
concentrates in a process similar to that used for producing zinc from 
zinc concentrates. Although the process has not yet been commercially 
adopted, it will warrant our consideration for a brief space. 

The principal problem in the preparation of copper concentrate for 
leaching is to find a preliminary treatment which will put the copper 
in a soluble form but still leave the bulk of the iron in such condition 
that it is insoluble in the solvent used , here the actual distribution of 
copper, iron, and sulfur in the compounds which make up the calcine 
is of paramount importance. The methods which have been in- 
vestigated aim to do this by (1) sulfating as much of the copper as 
possible to form the water-soluble CuS0 4 and (2) oxidizing as much 
of the iron as possible to Fe 2 3 , which is insoluble in water and in 
dilute sulfuric acid; if some copper remains as an oxide it can be 
leached with dilute acid. Ashcroft's 6 investigation was essentially 
for the purpose of determining if the requisite conditions could be at- 
tained by a controlled roast. We have already noted some of the 
facts which he has listed, and it appears that two of the conditions 
necessary for such a roast are: 

1. Preliminary roasting without nabbling to completely oxidize all 
free sulfides and form sul fates. 

2. A finishing roast in a rabbled apparatus at the proper temperature 
to decompose the sulfates of iron, but lea\e the copper sulfate 
unchanged. 

6 Sulphatizing Roasting, op. cit., p. 420. 



70 ROASTING 

In addition to these items, there are a number of other factors of 
importance which we shall not discuss in detail; for example the 
formation of copper ferrites must be avoided. These are insoluble in 
dilute acid, and would result in copper losses on leaching. The presence 
of ferrites in a calcine which is to be smelted, however, is not of great 
importance because these compounds decompose readily in the smelt- 
ing furnace. 

Another methocf for the differential sulfating of copper which has 
been investigated is the process of baking a previously roasted con- 
centrate with sulfuric acid. The preroasted calcine is mixed with the 
proper amount of sulfuric acid and then baked at the proper tempera- 
ture to cause the maximum formation of copper sulfate. 7 

The two principal objects which would be attained by a successful 
technique for preparing concentrates and then leaching them, arc: 
(1) the purified leach solution (copper sulfate solution) could be 
electrolyzed and the process would produce highly purified electrolytic 
copper directly and (2) such a treatment might well be less expensive 
than smelting followed by converting, fire refining, and electrolytic 
refining. On the other hand, in addition to the difficulties inherent in 
the process itself, the recovery of precious metals found in copper con- 
centrates must be considered. The matte formed in smelting is an 
excellent collector for precious metals, and these remain with the 
copper until separated by electrolytic refining. A sulfuric acid leach, 
however, would not remove the precious metals, and it leaching 
methods are ever to compete with smelting for the treatment of 
copper concentrate, there will have to be a parallel development 
of a suitable process for recovering the precious metals in the 
leach residue. 

In addition to the removal of sulfur from the roaster charge, arsenic 
and antimony are removed to some extent. Arsenides and antimonides 
found in the roaster feed are oxidized. The lower oxides As 2 0a and 
Sb 2 3 are quite volatile and pass off with the roaster gases. In an 
oxidizing atmosphere, however, much of the arsenic and antimony 
will oxidize to the higher oxides As 2 5 and Sb 2 5 , which are less vola- 
tile and form stable, non-volatile arsenates and antimonates with other 
metallic oxides. Usually it is necessary to alternate oxidation and 
reduction several times to completely remove arsenic and antimony, and 
in ordinary roasting operations only a part of these elements will be 

7 Floe C. F., and Hayward, C. R, Differential Production of Soluble Sulfates 
from Mixtures of Metallic Oxides: Am. Inst.^Mm & Met. Eng Tech. Pub. 735, 
1936; Floe, C. F, Extraction of Copper from Roasted Concentrates by Sulphuric 
Acid Baking Idem, Tech Pub. 768, 1937. 



NORANDA 71 

volatilized, the exact amount depending upon the nature of the at- 
mosphere within the roaster. 

SOME EXAMPLES OP ROASTING PRACTICE 

Copper Cliff. 8 At the Copper Cliff smelter of the International 
Nickel Company, the feed to the roasters is a nickel-copper concentrate 
containing considerable amounts of iron sulfidr-s. The plant contains 
30 Nichols Herreshoff roasters, which are located in two rows over the 
burner end of the reverberatory furnaces, with &ix roasters for each 
reverberatory furnace. 

Each roaster has a top drier hearth and 10 interior hearths; the 
height of the roasters is 31 feet 8 inches, and the outside diameter is 
21 feet 6 inches. The roaster shell is ^-inch plate lined with 9-inch 
firebrick and strengthened at each hearth with 8- by %-inch steel 
bands. The hearths are constructed of high-alumina firebrick. There 
are two rabble arms for each roasting hearth, and these are attached 
to the central shaft by a single pin; the arms can be easily removed 
through the furnace doors. All rabble arms on the ten interior 
hearths and the central shaft are air cooled; this cooling air may be 
allowed to escape into the furnace to provide air for combustion, or 
it may be conducted to the outside air from the top of the shaft. The 
rabble blades on the interior hearths are of white cast iron, and the 
blades for the drier hearth are of mild steel. The shaft rests on a 
step bearing and is operated by a 25-horsepower motor. The shaft 
is usually operated at a speed of 2 rpm, but a two-speed step pulley 
permits operation at 0.5 rpm when tins is desired. 

The maximum charge to each furnace is 275 tons per day, and the 
sulfur content is reduced from 28 per cent to 16 per cent. Oil heating 
is used only at times when the operation of the roasters has been 
interrupted; when operating normally the combustion of the sulfides 
supplies all the necessary heat. 

Noranda. 9 At the Noranda smelter, eight seven-hearth Wedge 
roasters are used to serve two reverberatory smelting furnaces. These 
roasters were originally designed to handle 100 tons of charge per 
day each, but by stepping up the speed of rotation from 0.75 rpm to 
1.09 rpm the capacity of each furnace was increased to a maximum of 
325 tons per clay; at this speed the sulfur content is reduced from 25 
per cent in the feed to about 11 or 12 per cent in the calcine. 

The material roasted averages about 65 per cent smelting ore, 19 

8 Canadian Min Jour., Vol 58, No 11, p. 673, 1937. 

Boggs, W. B , and Anderson, J N., The Noranda Smelter: Am. Inst. Min. <fe 
Met. Eng Trans., Vol. 106, p 183, 1933. 



72 ROASTING 

per cent fluxing ore, and only about 16 per cent concentrates; the 
copper content of the roaster feed is comparatively low, ranging from 
3.5 to 5 per cent copper. This charge is not self-roasting, so 5 to 10 
pounds of pulverized coal must be burned for each ton of calcine 
produced. The smelting ore, which makes up most of the roaster 
charge, is massive sulfide ore containing chalcopyrite, pyrite, and 
pyrrhotite; it is simply crushed and is not ground fine as is the 
concentrate. The fluxing ore is an acid flow rock mineralized with 
sulfides, and the concentrate consists almost entirely of chalcopyrite 
and pyrite. The relatively coarse size of the particles of crushed ore 
and the fact that a large quantity of pyrrhotite (which does not de- 
crepitate as does pyrite) is present tend to make the charge rather 
difficult to roast. The high sulfur and low copper content of the 
calcines means that the smelting of this calcine yields a low-grade 
matte (18 to 24 per cent copper), as the amount of sulfur eliminated 
in the reverberatory furnace by the interaction of sulfides and oxides 
(see Example 6, Chapter II) is not great. It is of interest to briefly 
consider some of the factors that are responsible for this particular 
roasting practice. 

1. The temperature of the calcine formed in a " dead " roast is 
always lower than the temperature of a " green " (incompletely) 
roasted calcine. This means that the dead roasted calcine would carry 
less sensible heat into the reverberatory, and more fuel would be 
needed for smelting. 

2. Roasting of high-iron ores and concentrates produces quantities 
of Fe 3 4 and Fe 2 3 in the calcines; these oxides must be reduced to 
FeO in the smelting furnace before they can unite with the silica to 
form a slag. The most important reactions for the reduction of these 
oxides are: 

3Fe 2 3 + FeS -> 7FeO + SO 2 
3Fe 3 4 + FeS -> lOFeO + S0 2 

If, therefore, the calcine is roasted to a low sulfur content, there is 
not sufficient sulfur left (as FeS) to reduce the higher oxides of iron, 
and as a result a low-sulfur calcine is more refractory toward smelting 
than a calcine containing more sulfur. Note that these reactions 
cause the elimination of a certain amount of sulfur from the reverbera- 
tory charge. 

3. At this plant the production of a low-grade matte means that 
more of the fluxing ore is needed in the converters, and the practice is 
controlled so as to keep the converters operating at full capacity. 
Thus the total smelting capacity of the plant is actually increased 



ANDES 73 

because of the additional amount of the fluxing ore which is consumed 
by the converters. 

Anaconda. 10 There are 14 seven-hearth roasting furnaces of the 
Anaconda McDougal-Wedge type, 25 feet in diameter, in the smelter 
at Anaconda, Montana. These each treat about 200 tons per day of a 
concentrate containing about 25 per cent copper and 32 per cent 
sulfur, bringing the sulfur content of the calcine down to about 18 
per cent. Small quantities of oil are used when necessary, but as a 
rule these concentrates are self-roasting 

It is interesting to compare these figures with those previous to 1927. 
Up to this time the roasters had a feed containing about 12 per cent 
copper and 36 per cent sulfur, and it was neo'js-ary to roast to about 
8 per cent sulfur in the calcine to produce the desired grade of matte. 
In 1927 the practice in the mill was changed so that the concentrate 
delivered to the roasters assayed about twice as much copper. This 
meant (1) that the tonnage of roaster feed was cut in half, (2) that 
the sulfur content of the feed was lower, and (3) that because of the 
higher copper content it was not necessary to roast to as low a sulfur 
content to get the same grade of matte. Previous to 1927, a single 
roaster would handle about 40 tons of charge per day, as compared with 
200 tons after roasting was started on the new concentrate. 

Andes. 11 The Andes plant includes a copper smelter for the 
treatment of sulfide concentrates and a leaching plant for low-grade 
oxidized ore. The sulfunc acid for the leaching plant is supplied from 
an acid plant which makes acid from the sulfurous gases obtained by 
roasting the sulfide concentrates There are two separate roasting 
plants the sulfide roaster plant and the acid-plant roasters. The 
concentrate to be roasted for acid-making must be high in sulfur and 
low in arsenic; consequently the mill produces two grades of concen- 
trate one low in arsenic and high in sulfur and another which con- 
tains the remainder of the copper and arsenic. This second concentrate 
is high enough in copper to be smelted directly after drying. 

The sulfide roasting plant contains seven Wedge roasters, each 
22 feet in diameter with seven roasting hearths and a drier hearth. 
Oil burners are used on the third, fifth, and seventh hearths; the oil 
is burned in firebrick muffles or combustion chambers set into one of 
the inspection doors on each of these hearths. These muffles are 2 feet 
long and have an opening 8 inches square. The roaster shaft makes 

10 Bender, L. V., Development of Copper Smelting a* Anaconda Eng and Min. 
Jour , Vol. 128, No. 8, p 301, 1929. 

n Callaway, L A and Koepel, F N, Metallurgical Plant of Andes Copper 
Mining Co.: Am. Inst. Min. & Met Eng Trans, Vol. 106, p 683, 1933. 



74 ROASTING 

0.75 rpm and the tonnage per roaster day ranges from 90 to 160 tons, 
depending upon the moisture content of the feed and product, 

While these roasters may be used for actual roasting they are usually 
used only as driers for both types of concentrate. The high-copper 
concentrate and flux is dried to about 3 5 per cent moisture and then 
goes directly to the reverberatory ; the high-sulfur concentrate is dried 
to below 1 per cent moisture, and it then goes to the acid-plant roasters. 
Fuel used for this drying ranges from 0.04 to 0.08 barrel of oil per ton 
of charge. 

There are seven Wedge roasters in the acid plant similar to those 
used in the sulfide roasting plant. The feed to these roasters is the 
dried concentrate from the sulfide roasters. The calcine goes to the 
reverberatory furnaces, and the gases are cleaned and conveyed to the 
Glover towers of the acid plant. 

Flin Flon. 12 Three Nichols Herreshoff roasting furnaces are em- 
ployed at the Flin Flon smelter of the Hudson Bay Mining and 
Smelting Company, Ltd., Flin Flon, Manitoba. Each roaster is 21 
feet 6 inches in diameter and 30 feet 3 inches high, outside dimensions, 
and contains 10 interior hearths and 1 drier hearth. The central 
column and rabble arms are air cooled, and the column turns at 2 rpm. 
Each roaster has a pulverized coal burner on the eighth hearth and an 
auxiliary burner on the fourth hearth; the auxiliary burner is used 
when the moisture content of the charge is higher than average. 
Additional heat is provided by electrically heating the combustion air 
which enters the bottom hearth of each furnace. This method is 
economically possible because of an adequate supply of cheap power. 
Each furnace has its own preheater which contains 18 ribbon heating 
elements over which the air is blown by a fan; the heated air passes 
into the roaster at 180 to 200 C. These preheaters operate at 600 
volts and have a rating of about 660 kw. 

The feed to the roasters consists principally of moist flotation con- 
centrate (15 per cent moisture on an average) plus smaller amounts 
of direct smelting ore and flux. The copper content of the roaster feed 
is about 7.35 per cent, iron 28.9 per cent, and sulfur 28.7 per cent. 
Sulfur is reduced to about 13 per cent in the calcine. The roasters 
can handle a maximum of 360 tons of charge per roaster day and con- 
sume 74 pounds of coal per ton of charge roasted less than half 
the amount of coal used before the air preheaters were installed. 

12 Ambrose, J. H., Flin Flon Copper Smelter: Canadian Min. Met. Bull. 281, 
September 1935. 



ROASTING 75 

SUMMARY 

The most important type of roasting in the metallurgy of copper is 
the roasting of sulfide concentrates in multiple-hearth roasters, the 
calcine from whi^h goes directly to reverberatory smelting furnaces. 
The primary purpose of roasting is to reduce the sulfur content to give 
the proper grade of matte when the calcine is smelted. The sensible 
heat in the calcine is utilized in the reverberatory furnace, and this 
reduces the amount of fuel required for smelting. 

As may be noted from the examples we have cited, the details of 
roasting practice vary considerably with the nature of the material to 
be roasted. Smelters treating high-grade copper concentrate may use 
the roasters simply for drying, or may dispense altogether with the 
roasting operation and send the cold, wet concentrate directly to the 
smelting furnace. 

Blast roasting and sintering are used to prepare sulfide ores and 
concentrates for blast furnace smelting; this method is not nearly as 
important as hearth roasting, however, because only a relatively small 
amount of copper-bearing material is smelted in blast furnaces. Some 
experimental work has been done on the flash roasting of copper con- 
centrates, but no commercial application has been made as yet. Studies 
have also been made of methods of using controlled roasts to render 
copper concentrates amenable to water or acid leaching. 

We have not considered in this chapter the gases produced in the 
roasting operation, the dust losses, nor the methods used in recovering 
the dust and treating the gases. After we have discussed smelting, 
converting, and refining we shall devote a separate chapter to a con- 
sideration of the gases and smokes produced in the pyrometallurgy of 
copper. 



CHAPTER IV 
SMELTING 

INTRODUCTION 

Copper smelting is a pyrometallurgical process in which solid ma- 
terial is melted and subjected to certain chemical changes. Products 
of a smelting furnace are liquids (slag, matte, metal, etc.), gases, and 
solid material carried out in the gas stream (dust and fume) The 
copper-bearing material to be smelted may be (1) ore, (2) calcines, 
(3) sinter, or (4) raw (unroasted) concentrates. Suitable fluxes are 
charged with the copper-bearing material to form a slag; the nature 
of the fluxes will be determined by the impurities in the charge. They 
may be either barren fluxes or revenue-bearing materials (copper ores, 
gold ores). Fuel used in blast furnaces is coke, and this coke is mixed 
with the rest of the charge; reverberatory smelting furnaces are fired 
with such fuels as oil, fuel gas, or pulverized coal. 

The copper may be tapped from the smelting furnace either as matte 
or as crude metallic copper; the smelting furnace may be either a re- 
verberatory furnace or a blast furnace In modern practice, by far 
the most important type of copper smelting is the smelting of either 
calcines or raw concentrates to copper matte in reverberatory furnaces, 
but there are other types of copper smelting which we shall consider 
also, viz.: 

1. Matte smelting in the blast furnace. 

2. Electric smelting for matte. 

3. Smelting of native copper concentrates in the reverberatory to 
produce metallic copper. 

4. Smelting of high-grade oxidized material to produce a crude 
metallic copper (" black copper"). 

REVERBERATORY MATTE SMELTING 

A reverberatory furnace is a long shallow furnace consisting of a 
hearth or laboratory, side and end walls, and a roof. The furnace is 
heated by means of burners placed in one end wall, and the products 
of combustion escape at the other end. A long-flame fuel is used 
gas, fuel oil, or pulverized coal and the flame extends over a large 

76 



DEVELOPMENT OF REVERBERATORY SMELTING 77 

part of the hearth. The material on the hearth is heated by radiation 
from the flame. The reverberatory is essentially a melting furnace, 
and there is ordinarily no extensive reaction between the gases in the 
furnace atmosphere and the charge on the hearth; it is possible to get 
some oxidation of the charge by using a large excess of air for the 
combustion, but this is wasteful of heat and is seldom practiced. As a 
rule the principal chemical reactions that takr place in the charge 
of a reverberatory furnace are reactions between various constituents 
of the charge itself; we have already noted some of these in Chapter II. 

Before presenting a description of the modern copper reverberatory 
it will be profitable to give brief consideration to the history and de- 
velopment of the reverberatory copper smelthig furnace in the United 
States. Much of the material given here is taken from an article by 
Frederick Laist 1 dealing with the development of the reverberatory 
furnace at Butte and Anaconda. 

Development of Reverberatory Smelting. In the early days of cop- 
per smelting in the United States, reverberatory smelting was practiced 
at Butte and Anaconda, Montana, almost from the beginning of the 
exploitation of the copper deposits of the district. Reverberatory 
furnaces were used in other parts of the country as well, in fact were 
used in Colorado some 30 years before the first furnace was constructed 
in Montana. It was in the Montana district, however, that the most 
intensive study of reverberatory smelting was made, and many of the 
improvements discovered there were instrumental in establishing the 
superiority of the reverberatory furnace over the blast furnace in the 
smelting of copper sulfide ores and concentrates. 

The first reverberatory furnace in the State of Montana was built 
at Butte in 1879 at the plant of the Colorado Smelting and Mining 
Company. This furnace had a hearth 14 feet long by 9 feet wide 
and used wood as fuel. It smelted about 10 tons of ore per 24 hours 
and produced a matte assaying 60 per cent copper and from 700 to 
800 ounces ot silver per ton. This matte was then hauled 200 miles to 
a railroad and shipped abroad This furnace and others built in the 
period 1879-1890 were similar to those used in Wales, which was up to 
then the world's foremost center of copper smelting and refining; similar 
furnaces had been operated in Wales for at least 100 years previous 
to this time. 

Figure 1 is a plan and section of a wood-burning matte furnace of the 
early 80's. These furnaces were fired by means of a firebox; the flame 
was drawn over the bridge wall between the firebox and the hearth, 

1 Laist, Frederick, History of Reverberatory Smelting in Montana, 1879 to 1933: 
Am. Inst. Min. & Met, Eng. Trans., Vol. 106, p. 23, 1933. 



78 



SMELTING 



and the products of combustion passed out the stack at the opposite 
end of the furnace. Roasting the copper ores was done in wood-fired 
hand-rabbled reverberatory roasters, and the calcine was cooled by 
quenching it in water; the calcine going to the reverberatories usually 




Plan 







. J 


i 








*"fi r-JTj~y 7t- 




\ 




: 'ta 


^, : "e? 1[ IJSkimmino Line^-^ 








1 


i: i ^ 






/ 


rP^j 








/ 


1* *J* 


J////// ' ' ' '" ''.,.'. ',/.'' '^ 


',"" 








\ I nneitudmal Section 







(Laist, Am Inst Mm & Met Ena Trans , Vol 106, p 26, 1933) 

FIG. 1. Wood-Burning Matte Furnace of the Early Eighties. 

contained about 10 per cent moisture About 3 tons of the wet calcine 
would be shoveled by hand into the furnace through the side doors, and 
after this had been melted down the slag would be skimmed, cast into 
slabs in sand beds, and then wheeled to the dump on barrows. After 
every third charge the matte was tapped into sand molds, cooled, sacked, 
and shipped abroad. The inside of the empty furnace was then 
examined, and wet crushed quartz was shoveled onto corroded portions 
of the hearth ; the furnace was then ready for another cycle. The fuel 
used (wood or coal) was also loaded or shoveled into the fireboxes 
by hand. 

The external shape of these early furnaces was rectangular, but the 
hearth itself was oval. Side walls were constructed of firebrick, and 
these walls supported the arched roof; the walls were lined with another 
layer of firebrick which could be replaced without disturbing the roof. 
The furnace was strongly braced by means of external steel I-beam 



DEVELOPMENT OF REVERBERATORY SMELTING 79 

buckstavcs and tiercels. The bridge wall between the firebox and 
hearth was built around a hollow steel " conkerplate " through which 
air was circulated ; this served the double purpose of strengthening the 
bridge wall and cooling it. The roof arch was made of silica brick, 
and it sloped rather steeply toward the front of the furnace; at this 
end the roof was usually only 12 to 14 inches above the level of the 
skimming plate. 

The hearth was constructed of quartz or sand containing about 97 
per cent silica; sometimes 3 or 4 per cent of crushed slag would be 
mixed with it to help sinter it into shape. After a new bottom had 
been placed in a furnace, the doors were closed and the furnace fired 
slowly until the maximum temperature was reached and then main- 
tamed thus for several hours. The furnace was then allowed to cool, 
crushed slag was spread over the surface, fused, and absorbed by the 
porous sand bottom. A second similar fusion completed the saturation 
of the lower hearth, and then more sand was thrown in to form the 
second hearth; this was saturated with slag and smelted in in much the 
same way as the bottom hearth It was essential that the hearth be 
prepared very carefully to give a solid monolithic mass; improperly pre- 
pared hearths would break up, and pieces would float to the top of the 
bath of matte during the regular smelting After the hearth was 
completed, the furnace was jcttlcd, and was then ready for its regular 
work. Sand or crushed quartz was heaped up along the side walls of 
the furnace to protect the bricks in the side wall from corrosion by 
the slag, and this process was known as fettling; the term fettling is 
also applied to the material used for this purpose. 

As far as principles are concerned, there is little or no difference 
between the methods used for smelting matte today, and those which 
were practiced in 1880 and 1890. The furnaces, however, have changed 
radically in the course of half a century; the capacity of the furnaces 
has increased fiftyfold, the heat required to smelt a ton of charge 
has decreased to about one-third of what it was originally, and the 
amount of necessary labor has been enormously reduced. Before 
turning our attention to the improvements which led to the modern 
furnaces, let us briefly summarize a few of the characteristics of these 
early furnaces. 

1. The early furnaces wore built over an open space, through which 
air was circulated to keep the hearth cool. It was believed that this 
was necessary to the satisfactory operation of the furnace. On this 
point Laist 2 states 

2 Laist, Frederick, op. cit , p. 34. 



80 SMELTING 

It is difficult to understand why that part of a reverberatory which is 
hardest to keep hot should have been deliberately cooled, but such had been 
the custom for generations and ideas firmly rooted in the past die hard. It 
is to the credit of the Montana metallurgists that they recognized the 
anomaly and had sufficient enterprise to break away from it. Gradually 
the modern practice of constructing the furnaces on a solid foundation 
became universal. 

2. Charging was done by hand ; matte and slag were tapped, cast, and 
transported by hand labor; and the firebox was fired by hand. These 
operations definitely limited the size of the furnace and contributed 
heavily to operating costs. 

3. The hearths of the furnace were oval in shape, and the roof sloped 
sharply downward near the front (flue end or " verb ") of the furnace. 
It was believed that this shape was essential to proper operation of 
the furnace, but it was later shown that this was not an important 
factor. 

4. Calcines were usually quenched in water and charged into the 
furnace wet. This not only dissipated the sensible heat in the hot 
calcines but made it necessary for the reverberatory to evaporate 
considerable water. 

5 The smelting process was essentially a batch operation; charges 
were added and smelted down, and after a sufficient amount of matte 
had collected (say after three charges), the slag was skimmed and all 
the matte was tapped out. It was believed that matte was harmful 
to the hearth and that the matte must all be tapped out at frequent 
intervals, the hearth patched up, and the furnace fettled. 

6. The charge was ordinarily high in copper and readily fusible; the 
matte fall was heavy, being 25 to 30 per cent of the total weight of the 
charges. The matte ordinarily contained from 50 to 65 per cent 
copper and considerable silver. These items permitted the use of 
technique which would be prohibitively expensive in smelting leaner 
and more refractory material. 

7. It was at first believed essential that each reverberatory furnace 
have its own stack; later this was found unnecessary, and it became 
common practice to connect several furnaces to a common stack by 
means of a system of flues. 

One of the first changes to be made in reverberatory smelting furnaces 
was to increase their size. It was evident that it would be more 
economical to operate a single large furnace than several small ones 
if the same amount of material could be smelted, and for a long 
period of time the size of the furnaces steadily increased. The length 
increased more rapidly than the _width, because the width was limited 



DEVELOPMENT OF REVERBERATORY SMELTING 81 

by the fact that the arched roof would not support itself over too wide 
a span. As the size of the furnaces increased, charging, tapping, and 
handling of material became more difficult, so that other innovations 
in furnace design nnd practice became necessary. 

Another factor that was of great importance was the ratio of the 
area of the grate in the firebox (which determined the rate at which 
fuel could be burned) to the area of the hearth. The firebox was 
lined with firebrick, and the grate was usually ordinary iron rods about 
an inch square; a pit was provided below the grate for the removal of 
ashes. In the small furnaces which were in use in 1880, the hearths 
measured about 10 by 15 feet and tl e fireboxes 4 by 5 feet, giving a 
ratio of approximately 1 to 5 for the areas. At first, the tendency was 
to increase the hearth area more rapidly than the grate area on the 
theory that a better utilization of heat would be obtained and that less 
coal would be used per ton smelted. Accordingly, the 35- by 14-foot 
furnaces of a later date had 5- by 8-foot fireboxes giving a ratio of 
1 to 9, and the longer furnaces which followed had a ratio of 1 to 18. 
In these latter furnaces, however, most of the smelting w r as done in 
about one-third of the hearth. 

In this connection it will be well to consider briefly what is meant by 
the capacity or smelting power of a furnace; obviously this quantity 
is measured by the tonnage of charge smelted per day, but there are 
several factors which determine this. The reverberatory copper 
matting furnace has two simple functions: (1) to melt down the 
charge and form liquid matte and slag at a temperature high enough 
to insure free-running slag, and (2) to provide sufficient space for 
these liquids to collect so that they have time and opportunity to 
separate cleanly into two layers and be tapped out of the furnace 
separately. The furnace should accomplish these two tasks with the 
consumption of the minimum amount of fuel. 

Size alone does not determine the smelting capacity of a furnace, 
nor does the amount of fuel burned. If the size is increased more 
material can be charged into the furnace, but unless the amount of 
fuel burned is increased, there will be little or no increase in smelting 
power. Consequently as furnaces become larger they must be equipped 
to burn larger amounts of fuel per unit of time; if the fuel consumption 
is insufficient, part of the interior volume of the large furnace is simply 
wasted space. The flame temperature is also of importance, because 
the heat for smelting must come from the gaseous products of com- 
bustion, and these can transmit their heat to the furnace charge only 
if they are hotter than the charge. Suppose, for example, that the 
furnace slag had to be discharged at J000 C; if the flame temperature 



82 SMELTING 

were only 100 above this figure, only a small portion of the total heat 
of combustion would be available for smelting, and a large excess of 
fuel would be consumed. The maximum temperature attainable with 
a given fuel requires that exactly the theoretical amount of air be used 
for combustion; if too little air is used, part of the fuel is unoxidized 
and its heating value is lost; if too much air is used, the excess dilutes 
the products of combustion and lowers the flame temperature. 

In 1890, a number of reverberatones 27 feet long were installed at 
Anaconda, and, for the first time, were fed with hot calcine direct 
from the roaster bins by means of feed hoppers set in the roof of the 
furnace. This was more efficient than the feeding of wet calcine, 
which had been the previous practice. 

Another important improvement in operating technique was the 
maintenance of a large pool of matte in the furnace at all times The 
old idea that the furnace must be emptied, repaired, and fettled at 
frequent intervals went into the discard when it was found that the 
matte was not injurious to the hearth, as had been thought, but actu- 
ally served to protect it. Therefore it became the practice to tap only 
part of the matte and slag at a time and to leave a pool of matte in 
the furnace at all times. This matte layer was from 12 to 24 inches 
deep and covered the entire hearth from one end to the other. This 
new technique had many very important effects on furnace operation. 

1. By skimming only part of the slag, tapping only part of the 
matte, and feeding the calcine in small amounts, the temperature of 
the furnace could be maintained more nearly constant at all times, and 
smelting became a continuous rather than a batch process. 

2. When high-grade matte was being made for shipping, it was not 
essential that the matte be tapped at any particular time. However, 
when the smelters: began to install converters to treat their own matte, it 
was necessary that a supply of hot liquid matte be available at all 
times to supply the converters. The pool of matte in the reverbera- 
tory served as a storage reservoir from which matte could be draun 
as needed by the converters. 

3 The molten, semi-metallic bath of matte was a good conductor of 
heat, and aided in conducting heat to the hearth; this made it easier 
to keep the hearth hot than when it was covered with a thick layer of 
a poorly conducting solid charge. 

4. Hot calcines from which SO 2 gas is still evolving will flow like 
water. When these calcines were charged onto the heavy liquid layer 
of matte, they would flow over the top of the liquid, and the charge 
would level itself off. The old practice of leveling the charge by means 
of hand-operated rabbles and spades thrust through the side doors of 



DEVELOPMENT OF REVERBERATORY SMELTING 83 

the furnace was abandoned, and the time and heat losses caused by 
hand leveling were no longer important. The furnaces could now be 
built with fewer doors and openings, and the air leakage and heat 
losses diminished; also the size of the furnaces could be further in- 
creased because they no longer depended upon the limitations of man 
power as a controlling factor. 

5 Fettling was done at longer intervals sometimes only once 
a month. 

6. The deep bath was particularly important when grate firing was 
still used. It was necessary to grate the fires at 4-hour intervals, and 
during this period most of the evolution of heat stopped and the labora- 
tory of the furnace cooled off considerably. Tho heat stored in the 
bath helped to maintain its temperature even though the space over 
the hearth became much cooler for a short time. 

7. The principal disadvantage to the " deep-bath " smelting tech- 
nique was the danger of a break-out in the furnace, the matte pool in 
a large furnace would weigh some 150 tons, and this could cause con- 
siderable damage. 

The " deep-bath " smelting method prevailed for a long time and is 
still used in many plants In other plants it has been j>uperseded by 
the 4< dry-hearth " technique, \\hich we will take up shortly. Before 
we proceed to take up the modern furnaces and methods, let us con- 
sider the furnace shown in plan and section in Figure 2 This was a 
112-foot furnace. Note the great increase in size as compared with 
Figure 1, also the difference m the shape of the hearth This was 
still a grate-fired furnace but coal was fed to the firebox by means of a 
four-chute coal hopper, and an a^h sluice was provided beneath the 
grates. Calcine \\as fed through the center of the roof by means of 
three calcine hoppers located near the back of the furnace. Two waste- 
heat boilers were employed to abstract part of the sensible heat in the 
flue gases leaving the furnaces. Waste-heat boilers are now standard 
equipment on almost all reverberatory matting furnaces. 

In the process of development of the reverberatory, many other 
modifications were used, but most of these have not survived. For 
instance, tilting furnaces were built, and also regenerative furnaces in 
which the combustion air was preheated. Let us now consider some 
of the changes which followed the type of furnace illustrated in Figure 
2, and this will lead us up to the present-day furnace. 

One of the most important developments was the change from grate 
firing to firing by means of fuel oil, pulverized coal, or gas The 
burners used for the combustion of these fuels were set in the back 
wall of the furnace, and the grates and bridge wall disappeared the 



84 



SMELTING 




DEVELOPMENT OF REVERBERATORY SMELTING 85 

laboratory or hearth now occupied the entire space within the furnace 
walls. Choice of fuel was largely dictated by the location of the plant 
and cost and availability of the fuel; all three fuels are still in use 
at different places, and apparently there is no significant difference in 
their smelting efficiencies as calculated on their calorific powers. It 
was believed for a long time that the fuel used in a reverberatory must 
necessarily burn with a luminous flame, because much of the heat that 
the bath receives comes from radiation from the flame, and it was 
felt that a non-luminous flame would have such a low radiating power 
that its smelting efficiency would be low. Experience at Anaconda, 3 
how r ever, has shown that natural gas has a smelting efficiency per heat 
unit equal to that of the pulverized coal previously used; pulverized 
coal burns with a highly luminous flame, but the gas flame is non- 
luminous to such an extent that one can see from one end of the 
furnace to the other while the gas is on. 

Coal-dust firing caused some trouble at first because ash and un- 
burned particles fell on the bath and formed an insulating blanket. 
This had the effect of preventing the heat of the flame from reaching 
the bath, and the bath became chilled. It was found later, however, 
that this difficulty was caused by insufficient pulverizing of the coal 
arid when the coal was sufficiently fine this blanket of ash did not 
form. When firing with coal dust, part of the ash is carried out the 
flue; the rest falls on the charge and eventually becomes part of the 
slag and must be taken into account in calculating the amount of 
slag-forming fluxes to be used. Fuel oil and gas, of course, contain 
no ash. 

Another improvement in furnace design was the use of water-cooled 
side plates placed in the side walls of the reverberatory; the cross- 
section in Figure 3 illustrates how these side plates are located. 
These plates cool the side walls and hence prevent corrosion by the 
slag and the danger of the matte breaking out through the furnace 
walls. A system such as this serves to warn the operator if the walls 
are becoming thin, because then the water flowing out of the cooling 
plates is abnormally hot. At Anaconda it was found that the use of 
these side plates practically eliminated all trouble with matte break- 
outs and helped keep the furnace in better shape. 

The method of charging a reverberatory matting furnace has been 
the subject of much research, and a perfectly satisfactory method of 
charging has never been found. Since hand charging through side 
doors on the early reverberatories was abandoned, it has been standard 
practice to charge through openings in the roof; one important devia- 

3 Laist, Frederick, op c it., p. 87. 



86 



SMELTING 



tion from this has been the " gun feed " method of charging through the 
side walls. This is a recent development, and we shall consider it a 
little later. At first the charge hoppers were set near the center of the 
furnace roof, as shown in Figure 2, but later this method of center 
charging gave way to side charging, in which the calcine hoppers dis- 




Scale 



(Laist, Am Inal Mm d Met Eno Trana , Vol 106, p 80, 1VSS) 

FIG. 3. Cross-Section of Anaconda Center-Charged, Water-Cooled Reverberatory 

Furnace of 1928. 

charged through the roof close to the side walls. Thus the furnace 
charge piled up along the side walls, and this charge acted as its own 
fettling protecting the walls from corrosion by the slag bath and 
helping to prevent break-outs. About 1924 the copper-smelting com- 
panies were confronted with serious litigation involving the right to use 
the method of side charging which had been previously patented. 
During the trial of the Carson case, the Anaconda company made some 
investigations relative to certain controversial points, and these served 
to throw a good deal of light on the general subject of the operation 
of reverberatory smelting furnaces. When side charging was first used, 
it was found that it was no longer possible to keep a deep pool of 
matte over the entire hearth the matte pool froze over near the 
back of the furnace, and only a relatively small pool of matte was 
maintained near the front of the furnace. Although it had been 
thought that a deep bath was essential, it soon became evident that 
the " dry -hearth " technique was just as efficient as the previous prac- 
tice. In studying the merits of side versus center charging, then, it 
was also necessary to study the effect of deep-bath versus dry-hearth 



THE REVERBERATORY FURNACE CHARGE 87 

smelting, as both changes had been made at the same time. Several 
reverberatory furnaces were operated for a long time under different 
conditions, and the conclusion was reached that none of these factors 
had any appreciable effect on the smelting power of the furnace. 4 
In other words, the important thing in reverberatory smelting is to 
maintain a constant evolution of heat and keep plenty of unsmelted 
charge in the furnace so that the flame and hot gases have something 
to work on at all times; otherwise it makes little difference how the 
furnace is charged or whether a pool of matte is maintained in it or 
not. Charging in different ways, and smelting with or without a deep 
bath of matte, may affect such things as the life of the furnace, dust 
losses, convenience in manipulation, and metal looses in slag, but they 
have no significant effect upon the actual smelting power of the furnace. 
The Reverberatory Furnace Charge. The material fed into a re- 
verberatory furnace will depend upon the type of ore, the nature of the 
concentrate produced from it, and the amount of preliminary treatment 
(drying and roasting) that it has received. The important facts to be 
considered with respect to the solid materials charged are: 

1. The copper and sulfur content of the charge. This determines 
the grade and amount of the matte formed. 

2. The nature of the gangue or waste materials in the ore. These 
must be fluxed and slagged off, and the amount and nature of this 
gangue material determines the amount of flux that must be used and 
the amount or volume of slag formed. 

3. Whether the copper-bearing material has been roasted or not. 
In modern plants we find reverberatory furnaces operating on 
(a) roasted calcine, (b) dried concentrate, and (r) wet concentrates 
just as they come from the mill filters. 

4. The particle size of the solid material. 

5. The physical and chemical properties of the flux used. 

In addition to the solid material fed to the reverberatories, in most 
plants it is also necessary to treat the slag from the converters. This 
is a ferrous silicate slag, high in iron, and containing about 4 per cent 
copper; this comes directly from the converters and is charged into 
the reverberatories in the liquid form. The reverberatory must also 
handle a certain amount of reverts collected dust and fume, ladle 
skulls, and refinery slags. 

Another material which finds its way to some smelters is cement 
copper, a finely divided high-grade precipitate of metallic copper 
obtained by the precipitation or cementing of copper from copper sul- 



4 Laist, Frederick, op. cit., p. 78. 



88 SMELTING 

fate solutions on metallic iron. When cement copper is being treated 
it usually forms part of the reverberatory charge. 

Modern trends in concentration involve regrinding of middlings 
and concentrates with the subsequent production of finely divided 
high-grade flotation concentrate; this material presents many problems 
to the smelter. On the whole the furnace feed is becoming more basic 
as improved concentration decreases the amount of silica in the con- 
centrates; converter slag was formerly useful as a flux because of its 
high iron content, but this is generally no longer true. The increasing 
basicity of furnace charges is largely responsible for the fact that 
siliceous refractories are being replaced by basic refractories in many 
reverberatory furnaces. 

Roasting of finely divided concentrate produces a " wild " calcine 
which is difficult to handle without serious dust losses; these losses occur 
in the roaster itself, in transporting the calcine to the reverberatory 
furnace, and in charging the material into the furnace. Dust losses are 
particularly high when this " wild " calcine is charged through the roof 
of the furnace. 

The Smelting Action of the Reverberatory. The principal function 
of the reverberatory, as we have noted, is to melt down the charge and 
permit the liquid slag and matte to segregate into two layeFs. In 
Chapter II we have considered the chemical reactions which <4ake place 
during smelting; these are of two important types. _ * 

1. Formation of matte and slag by metathesis (double decomposi- 
tion) . Because of its strong affinity for sulfur, the copper on the charge 
forms Cu 2 S, the stable copper sulfide. This copper may enter the 
furnace as a sulfide in concentrate or calcine; as an oxide in calcine, 
oxidized ore, or converter slag; or as metallic copper in cement copper 
all eventually becoming sulfidized and entering the matte as Cu 2 S. 
The rest of the sulfur on the charge i* either volatilized as SO 2 or 
combines with iron to form FeS; the resultant liquid FeS is miscible 
with CuoS, and the solution of these two sulfides is the principal con- 
stituent of the matte. 

2. Reactions which result in the formation of SO 2 and the consequent 
elimination of part of the sulfur on the charge. Some sulfur is 
eliminated by the direct action of the flame gases on the piles of 
charge, but this roasting action in the reverberatory is not great be- 
cause there is usually not much excess air in the flame gases and because 
the material is not being stirred or rabbled. Other reactions which 
serve to eliminate sulfur by the interaction of constituents within the 
charge we have considered in Chapter II, Examples 4, 5, and 6. Sulfur 
elimination in the reverberatory depends upon the nature of the charge 



CONSTRUCTION OF THE REVERBERATORY FURNACE 89 

and the amount of free oxygen in the flame gases. This elimination 
may range from practically nothing up to 30 per cent of the total sulfur 
on the charge. The sulfur elimination by the furnace must be known 
in order to calculate the grade of matte and the matte fall to be ex- 
pected in any particular ore. 

In many reverberatory furnaces infusible accretions are formed which 
tend to build up on the hearth of the furnace The most prevalent 
of these is magnetite, Fc r{ 0, The magnetite may be already present 
in the calcine or it may be formed by the partial reduction of Fe 2 3 , 
as for example 

9F' 2 3 + FeS -* OFe 3 <) 4 + SO 2 4 FeO 

The first stage in the reduction of Fe 2 O ;; takes place readily, but the 
Fe : j() 4 formed is highly refractory since it is both difficult to melt and 
to reduce to FeO in which form the iron can be slagged. Magnetite 
may be deposited by simply settling to the furnace bottom, or some of 
it may dissolve in the matte and then precipitate on the hearth after 
the bath has become saturated with it. Tt seems that liquid matte has 
some solvent action for magnetite because magnetite crystals have 
been found in frozen mattes However the deposit may be formed, 
there is a gradual building up of the layer of magnetite on the hearth 
of most operating furnaces, and very often it is this that determines 
the length of the life (cnmpmqn} oi the furnace, because when the 
accretions become too thick the furnace must be shut down and the 
hearth rebuilt. 

The greatest depth of these magnetite accretions is found in the 
back or smelting zone of the furnace, and in " dry-hearth " smelting 
the solid charge rests directly on this " magnetite hearth." A pool of 
matte And slag forms near tlve frout of the furnace, but there is no 
dee}) bath of matte in the smelting zone Small pools of matte and 
slag accumulate on this '* plateau " in the smelting zone, from which 
the liquids trickle do\\n to the collecting pool near the front of the 
furnace. A cross-section of a coal-fired dry-hearth reverberatory is 
illustrated in Figure 4, and two of the diagrams in Figure 5 show the 
" magnetite hearth." The matte pool in this particular furrace held 
about 50 tons of matte; in a similar furnace operated with a deep bath 
of matte, the matte pool contained about 200 tons. 

Construction of the Reverberatory Furnace. As we shall see when 
we consider some typical reverberatory furnaces, there are many differ- 
ences in details of their construction. At this point we shall consider 
some of the general characteristics. 

Furnaces are usually constructed of silica brick with a monolithic 



90 



SMELTING 



t. 




silica bottom; this bottom is built up 
over several courses of silica brick 
which in turn rest upon a solid founda- 
tion of concrete or poured slag. Steel 
buckstaves (usually I-beams) rise ver- 
tically along the sides of the furnace, 
and their lower ends are set firmly in 
the foundation. Steel tierods connect 
the buckstaves across the top of the 
furnace, and this combination of buck- 
staves and tierods holds the furnace 
together. Water-cooled hollow metal 
plates may be set in the side walls to 
cool and protect the refractory brick 
and keep the steelwork from buckling 
under the prolonged heating. 

The Bottom. The furnace bottom or 
smelting hearth is usually constructed 
of silica sand or crushed quartz, formed 
into shape and sintered by the heat 
of the furnace; usually some slag or 
matte is added to help sinter the hearth. 
The liquid slag and matte in an oper- 
ating furnace seep into the hearth, and 
magnetite also collects on the hearth; 
these substances replace the original 
material, and it is usually found on 
shutting down a furnace that the orig- 
inal hearth material has been com- 
pletely replaced by matte, slag, and 
magnetite. Old hearth material is re- 
moved, crushed, and resmelted to re- 
cover the contained copper. 

The Side Walls. Reverberatory fur- 
nace side walls are constructed of 
silica brick; they are usually quite 
thick near the bottom of the furnace 
and thinner near the top. Some fur- 
naces have one or two courses of mag- 
nesite brick laid on the inner sides of 
the walls extending from the hearth to 
a level above the slag line. This mag- 



CONSTRUCTION OF THE REVERBERATORY FURNACE 



91 





p ,/. - :: -; *^ ' . / '.-';.;.'." 


g^z-- 


Hearth 


x^HS] 

Ttrjr ,: 


Silica Bottom 


H J 1 

J-r-jJ'-i-l ! 



j^i^xrTz^ Ci; ^ 




(Loi(, /1m. Jrw<. Afin. A Afct. Eng. Tmn^ , Vol. 106, p. 7^, 

Fio. 5. Keverberatory Cross-Sections. 

a, Showa " dry hearth " and charge piles resting on solid " magnetite " hearth. Side-charging. 

b, HhowB center-ohartcing with " dry hearth ", charge resting on solid " magnetite " hearth, 
r, Center-charged, " deep-bath " furnace, charge floating on molten slag and matte bath. 



92 



SMELTING 




FIG. 6. Side View of Reverberatory Matting Furnace Looking Toward 
the Firing End. 

In the background is the launder for charging converter slag into the furnace. Note the buckstaves, 
skewback plate, and tieroda; also the panels in the side walls. Theae panels allow more brick to be 
added on the outside when the original brick has burned out, 

nesite lining may not extend for the entire length of the furnace but 
may only line the crucible or the front portion of the furnace. When 
a new furnace is made ready for the first charge the side walls are 
fettled, and more fettling material can be added as needed by means of 
charging pipes near the side walls. In side-charged furnaces the 
charge tends to act as its own fettling material. 

The Roof. The roof or arch of a reverberatory furnace is made of a 
single thickness of special bricks. Part of these bricks may be 
" straights/' and part or all of the bricks are wedge shaped. These 
are set with the widest part on top so that they form a shallow arch. 
The arch sets against the skewback plates which run along both sides 
of the furnace and are fastened to the inner sides of the buckstaves. 
The weight of the arch rests on the side walls, and the thrust of the 
arch is taken by the skewback plates. Thus the arch is self-supporting 
and the width of the furnace is largely governed by the working strength 
of the arch brick at furnace temperatures. The maximum span which 
will support its own weight is about 30 feet; consequently the width 
of a furnace cannot be much more than 30 feet. The roof often slopes 
downward near the front of the furnace. 



CONSTRUCTION OF THE REVERBERATORY FURNACE 93 

Sprung arches, such as have been described, are usually made of 
silica brick; their thickness may range from 9 to 20 inches. Recently 
some furnaces have been constructed with part or all of the arch made 
of magnesite brick. Magnesite brick arches are of the suspended type; 
special shapes of brick are used which hang from supporting rods 
above the furnace (Fig. 8). Thus the weight of the roof is carried 
by external support, and the thrust on the skewback plates does not 
hold the arch in shape as is the case with a sprung arch. Magnesite 




FIG. 7. Interior of a Reverberatory Furnace Looking Toward the Skim End. 

Thia furnace haa a sprung arch roof and the fettling is in place along the aide walla. 

brick is denser than silica brick and shows less strength at high 
temperatures; so far it has not been feasible to build non-suspended 
(sprung) arches of magnesite brick over spans as wide as those used 
in these furnaces. 

All refractories expand when heated, and in designing a furnace, 
allowance must be made for this expansion as the furnace comes to 
temperature. Usually the longitudinal expansion of the roof is taken 
up by means' of suitable expansion joints gaps which close up as the 
roof expands. These expansion joints show as breaks in a cross-section 
and may be noted in Figure 9. Lateral expansion may be taken care 
of by adjusting the tension in the tierods which cross the top of the 
furnace and connect the buckstaves on opposite sides. 



94 SMELTING 

Refractories. For many years silica brick was the standard re- 
fractory used in the construction of furnaces. Bottom, side walls, and 
arch were all made of this material. Today there are still many 
reverberatories which are of silica brick construction throughout. In 
other plants, however, basic magnesite refractories are being used for 
crucible linings and roof arches. In modern practice it is necessary 
to treat charges which are increasingly basic in their chemical compo- 
sition and finer in size; these materials produce quantities of dust, 
and this basic dust (largely oxides of iron) reacts with the hot 
silica (acid) refractory to form fusible silicates. It is this corrosive 
action of basic material on siliceous refractories that has led to the 
use of basic refractories; these, of course, are not attacked by basic 
oxides. 

Spallmg, or the breaking of refractory material when subjected to 
sudden temperature changes or " thermal shock," is important because 
both magnesite brick and silica brick have rather strong tendencies 
to fail when subjected to sudden temperature changes; probably mag- 
nesite brick is the worse offender in this respect. Damage due to 
spalling may be minimized by careful operation of the furnace and the 
use of well made refractory brick. 

Chrome brick is occasionally used in portions of the furnace sub- 
jected to severe corrosion by matte, slag, and dust. This refractory 
is highly resistant to corrosion; its principal disadvantages are its 
high cost and the fact that it will absorb matte and form a mass which 
is very difficult to smelt to recover the absorbed copper. 

In an article on the application of refractories to the copper industry, 
Suydam 5 lists the following important conclusions: 

1. The trend in the choice of refractories has been toward the more basic 
types as smelter feed has become more highly concentrated. 

2. Fine grinding incident to flotation, and higher temperatures necessary 
to lower costs have imposed successively harder conditions, especially in the 
ore-melting reverberatory. 

3. The tendency to increase furnace widths and temperature of firing 
makes almost imperative the selection of a refractory better suited to the 
service than is the commonly used silica brick. 

4. Of known refractories, magnesite brick seem to possess the most 
desirable properties. Their high expansion characteristic, weight and cost 
have prompted careful increases in the proportion of magnesite substituted 
for silica in sprung arches of large span. 

5. Nothing was known of the behavior of magnesite brick in suspended 

6 Suydam, A. G., Application of Refractories to the Copper Industry: Am. Inat. 
Min & Met. Eng. Trans., Vol. 106, p 277, 1933. 



CONSTRUCTION OF THE REVERBERATORY FURNACE 



95 



construction. Such a roof was provided over an experimental furnace and 
data gathered regarding the following points: period of safe heating and 
cooling; temperatures at which cold shapes could safely be installed; tem- 
perature gradient through brick; probable heat losses through the roof; 
over-all expansion; the effect of commonly used expansion joints and some 
information on the probable causes of spalling. 

6. The probability of magnesite brick serving as a complete sprung arch 
over wide spans is considered in the light of such data as are available. 
Ideas as to the strength of magnesite brick at high temperatures may have 
been clouded by the results of tests under soaking heat conditions without 
due consideration of the mechanical strength of the refractory as actually 
used in practice, where a rather steep temperature gradient is had except 
in bottoms. 




Section on C L of Furnace 

REVERBERATORY FURNACE 
Hudson Bay Mining & Smelting Co Limited 

Present Construction 



FeetO 



5 10 15 20 25 30 

(Ambrose, Canadian M\n Met Bull. 281, p 410, 1936} 



Fia. 8. Reverberatory Furnace Showing a Suspended Magnesite Arch. 

The life of siliceous refractories in reverberatory furnaces may be 
greatly extended by the method of " hot-patching " developed at the 
Clarkdale smelter. As described by Kuzell, 6 this method consists in 
spraying a water suspension of refractory material on the surface while 
the furnace is under full fire. The refractory material consists of 
quartz-sandstone pulverized to 78 per cent irinus 200 mesh, and a 

8 Kuzell, C. R., Clarkdale Method of Hot-Patching Operating Furnaces: Am. 
lost. Min. & Mot Eng. Tech Paper 995 (Metals Technology), February 1939. 



96 SMELTING 

slurry of pulverized clay which is treated with live steam for several 
hours to insure complete dispersion and disintegration. The proper 
amounts of quartz and clay slurry are mixed in a concrete mixer, and 
the mix is then blown on to the refractory surface by means of a 
spray gun using a 1%-inch iron discharge pipe. An air pressure of 
50 to 60 pounds is used on the spray gun when patching arches; for 
side wall and flue repairs lower pressures are used. The mix must be 
applied in successive thin layers, allowing sufficient time between ap- 
plications to permit the heat of the furnace to set the refractory; an 
experienced operator can build a patch up to 6 inches thick. Since 
the adoption of this method at Clarkdale the furnace campaigns have 
been almost indefinitely extended, and when a reverberatory is shut 
down it is usually for external reasons and not because of failure of 
the refractories. 

With respect to smelting practice on the North American Continent, 
the following excerpts from the annual review number of the Febru- 
ary 1940 issue of Engineering and Mining Journal 7 are pertinent. 

Several differences were noticed between reverberatory furnace practice 
in the Southwest as compared with practice in Canadian plants. In the 
American plants, sprung arches of silica brick are used exclusively, whereas 
in Canada all copper and copper-nickel reverberatory furnaces have sus- 
pended magnesite roofs in the smelting zone at least, and many have 
suspended roofs for the full length of the furnace. 

A second notable difference is that in most of the American plants the 
cross-sectional area of the furnaces is uniform throughout their length, 
whereas in most furnaces in Canada the height of the roof at the firing end 
is increased to provide more area for combustion in the smelting zone, the 
cross-sectional area being decreased toward the front end. 

Tapping of Slag and Matte. The withdrawal of matte is usually 
called " tapping/' and the matte is removed through " tapholes " The 
removal of slag is known as " skimming," and the slag flows through 
a " skimming door." The term " skimming " originally meant the 
complete removal of slag preparatory to the tapping of the entire 
pool of matte; the term has lost its original meaning, as a certain 
amount of molten slag is retained at all times in modern furnaces 
even in dry -hearth smelting there is always a good-sized pool of matte 
and slag in the front end of the furnace. The skimming door is a 
small opening in the wall of the furnace; it is closed by means of a 
clay dam when the flow of slag is to be stopped. In some operations, 

7 Boggs, W. B., Copper Metallurgy: Eng. and Min. Jour., Vol. 141, No. 2, p. 88, 
1940. 



CONSTRUCTION OF THE REVERBERATORY FURNACE 



97 




98 SMELTING 

slag is skimmed intermittently by breaking and rebuilding the clay dam 
as occasion requires ; in other operations where a large volume of slag 
is produced the slag flows continuously through the skimming door as 
long as the furnace is in operation. 

The matte tapholes are located some 10 to 20 inches below the level 
of the skimming door; these are usually holes about 3 inches in diameter 
extending through the furnace wall; often there are two tapholes at 
different levels. These tapholes may be drilled through a refractory 
brick which is set in a metal plate in the furnace wall these serve to 
cool the refractory and prevent corrosion of the taphole. The tapholes 
are closed by ramming a clay plug into the hole, and the furnace heat 
burns the clay into a hard mass; when the matte is to be tapped, a 
steel tapping bar is driven through the hole to open it. 

Skimming doors and tapholes may be located in the front wall of 
the furnace or in either side wall ; the exact location is often determined 
by the plant lay-out, and the skimming doors and tapholes are placed 
in such a position as to provide the maximum convenience in the sub- 
sequent disposal of the furnace products. Usually they are located 
in the front wall, or in the side walls near the front of the furnace, but 
when deep-bath smelting is used it is possible to tap matte through holes 
set in the side walls near the firing end of the furnace. It is sometimes 
advantageous to have the skimming door in the side wall rather than 
in the front wall. This is because there is always a certain amount of 
material dripping from the verb arch near the uptake where the furnace 
gases enter the flue; these drippings are formed by reaction of the dust 
from the charge with the silica of the refractory and contain con- 
siderable copper. If the slag is skimmed at the point beneath the 
verb arch, the slag will be contaminated with these drippings and the 
copper loss in the slag will be higher. 

Wet-Charge Smelting. For the past 12 to 15 years the method of 
charging wet concentrates directly into the reverberatory furnace has 
been gradually developing, and today it is standard practice at several 
copper smelters. 

Following the success of wet-charge feeding as worked out by A. D. 
Wilkinson at Cananea, it was decided in 1927 to try out this charging 
method at the International Plant at Miami, Arizona. Practice at 
the Miami plant is given in a paper by Honeyman 8 from which the 
following discussion is taken. 

8 Honeyman, P. D. L, Reverberatory Smelting of Raw Concentrates at the 
International Smelter, Miami, Arizona: Am. Inst. Min. & Met. Eng. Trans., Vol. 
106, p. 88, 1933. 



WET-CHARGE SMELTING 




Pi, 10. cm 




FIG. 11. Tapping Slag from a Reverberatory Furnace. 



100 SMELTING 

Figure 12 shows the plan and section of one of the Miami furnaces 
equipped for wet charging. The furnace feed is stored in charge bins 
under the feed floor, and these feed by way of a pan feeder to two drag 
chain conveyors which run above the charge pipes that pass through 
the arch near the side walls. These charge pipes are 8 inches in 
diameter and are spaced at 44-inch intervals; they are closed by means 
of gates which are controlled from the charging runway. The charge 
pipes cover the entire charging zone, which extends for 65 feet down 
the furnace. In charging the furnace all the slides covering the 
charge pipes are opened, and the conveyor is started; the operator 
starts at the firing end and feeds the charge into two or three holes until 
they are filled up, then he passes along to the next series of holes, and 
so on until he has covered the entire charging zone. Each charge pipe 
is equipped with a peephole through which the condition of the charge 
within the furnace may be observed; it is usually necessary for the 
charger to insert a short rod through the peephole and assist the flow 
of the charge into the furnace. The amount of material added in 
one charge will often exceed 60,000 pounds, and the charging ordinarily 
takes from 20 to 30 minutes. Charging is done about six times per 
8-hour shift; the frequency and size of the charges will depend upon the 
nature of the charge. 

Deep piles of charge are kept along the sides of the furnace in the 
charging zone at all times, and a pool of slag and matte is maintained 
between these charge piles. Note that a band of magnesite brick is 
set in the side walls to protect the slag line, extending through that 
part of the furnace where there are no charge piles to protect the walls. 
Water will react with hot matte with explosive violence, but it is found 
that there is little or no shooting caused by the contact of the wet 
charge with the matte, provided that a sufficient depth of molten slag 
is maintained above the matte pool. As a matter of fact, it has been 
found that matte can safely be tapped from the furnace directly under 
the charge piles, and the diagram shows a spare taphole located well 
back in the smelting zone. The taphole in the front of the furnace, 
however, is the only one that is used as a rule. 

The average charge will contain 77 per cent flotation concentrates, 
4 per cent cement copper, 7.5 per cent flux, and 11.5 per cent plant 
secondaries or reverts. Copper will run about 33 per cent, and sulfur 
about 25 per cent; moisture will average 11 per cent, and at times 
reach 15 per cent. Molten converter slag is returned to the furnace 
through a launder discharging through the middle of the bridge wall. 
The converter slag often reacts with the bath to produce a violent boil- 
ing action which may extend throughout the smelting zone; this assists 



WET-CHARGE SMELTING 



101 




I 



to 

I 

1 

VI 



I 

03 

e 



I 



102 



SMELTING 



in mixing the surface charge with the molten bath and helps promote 
the smelting of the charge. 

Another paper by Leonard Larson 9 describes the wet-smelting prac- 
tice at McGill, Nevada. Figure 13 is a section through one of the 
reverberatory furnaces at McGill. Note that the walls and arch are of 
silica brick, and there is a shelf of magnesite brick set inside the side 




(Larson, Am. Inst. Mm. & Met. Eng. Tech. Paper 981, Metals Technology, Oct 1938) 
FIG. 13. Section Through Wet-Charged Reverberatory at McGill, Nevada. 

walls. This shelf runs for the entire length of the furnace, and extehds 
above the slag line. The raw charge is dropped through the roof of 
the furnace onto this shelf which supports the charge along the side 
walls, and prevents undue sloughing of the wet charge into the fur- 
nace bath. There are 24 charge hoppers on each side of the furnace, 
and the charge holes extend through the arch near the side walls. The 
charge is brought to the charge hoppers by means of two Traylor vi- 
brating conveyors ; these conveyors are approximately 80 feet long, and 
are simply steel troughs which are vibrated longitudinally, the vibra- 
tion causing the material to " crawl " through the trough. These 
troughs are equipped with spring gates so that the charge can be fed 
into any one of the charge hoppers; the hoppers have gates which are 
closed when not in use to prevent air entering the furnace. The ma- 

9 Larson, Leonard, Copper-Smelting Plant Remodeled for Direct Smelting: Am. 
Inst Min. & Met. Eng. Tech. Paper 981 (Metals Technology), October 1938. 



WET-CHARGE SMELTING 103 

terial from the reverberatory storage bins is brought to the vibrating 
conveyors by means of belt feeders. 

The feed to the furnace averages about 8 to 9 per cent moisture; 
with both conveyors operating, about 100 tons of this material can be 
charged per hour. These vibrating conveyors have been in service 
since October 1934 and have proved entirely satisfactory for the feeding 
of wet charge to the reverberatory furnace. 

At the new copper smelter of the Chino Copper Company at Hurley, 
New Mexico, blown in on May 2, 1939, wet-charge smelting is em- 
ployed. Here, also, the material is charged by means of vibrating 
conveyors above the side walls of the furnace. The crucible of this 
furnace is lined with magnesite brick, and two sections of the roof 
near the gas outlet are made of firebrick; otherwise the furnace is made 
of silica brick throughout. 10 

The advantages of wet-smelting practices may be listed as follows: 

1. The roasting plant is eliminated. This reduces the amount of 
equipment that must be maintained and the amount of handling neces- 
sary. Dust losses arc greatly diminished, and this improves the 
cleanliness and the general working conditions of the plant. 

2. Dusting within the furnace itself is greatly diminished, and con- 
sequently there is less wear on the refractories and less material to be 
caught in the dust-collecting apparatus and returned to the furnace. 
At Miami 11 during the period of calcine smelting, there was consider- 
able dusting, and under these conditions a furnace seldom operated 
over 9 months without a shutdown for general repairs. After the 
adoption of wet-charging methods the furnace campaign has been 
extended to well over 2 years. 

3. The accumulation of magnetite on the hearth is much less in a 
wet-charged furnace than in one smelting calcines. Probably the 
main reason for this is the absence of oxides of iron in the wet charges, 
as compared with the large amounts of Fe 2 3 found-in most calcines. 

The disadvantages of wet-charge smelting are as follows: 

1. The most obvious disadvantage, of course, is the fact that wet- 
charge smelting can be used only on relatively high grade feed. Many 
ores are difficult or impossible to concentrate sufficiently to be smelted 
directly to yield a satisfactory matte. For low-grade material it is 
better to eliminate some of the sulfur in the roasters so that the 
reverberatory will produce a higher-grade matte. 

2. The wet-charged furnaces operating at present in southwestern 
United States, Mexico, and Africa are all located in climates where 

10 Huttl, J*B., Chino Today: Eng. and Min. Jour., Vol. 140, No. 9, p. 29, 1939. 
n Honeyman, P. D. I., op. cit. 



104 



SMELTING 



there is little or no cold weather. Handling and feeding of wet con- 
centrate during the winter in such places as Montana and Canada 
would probably offer many difficulties. 

3. The fuel consumption per ton of charge smelted is, of course, 
greater when smelting cold, wet concentrate, than when smelting dry, 
hot calcine. If the concentrate were roasted, the burning of sulfides 
would provide some or all of the heat necessary to dry and heat the 
calcines, and less total fuel would be needed. 

Laist 12 estimates that if wet charging were to be used at Anaconda 
to replace calcine smelting, it would require a 50 per cent increase in 
the amount of fuel used. 




L Pane I Construction 

Section A-A n ~H Sectional Elevation 

(Wagstaff, Am. Inst Min. & Met. Eng. Trans , Vol. 106, p. 102, 19S3) 

FIG. 14. Principal Features of Gun Feeder and Construction of Side Wall. 

Gun-Feed Furnaces. The following discussion is taken from a paper 
describing the gun-feed method of charging a reverberatory furnace 
as developed at Garfield, Utah. 13 A sketch of the equipment used 
is shown in Figure 14. This feeder is used for charging a furnace with 
hot calcine. 

The gun feeder was designed to introduce the finely divided hot cal- 

14 Laist, Frederick, op cit., p. 81. 

13 Wagstaff, R. A., Development of Gun-Feed Reverberatory Furnaces at Gar- 
field Plant of American Smelting and Refining Co.: Am. Inst. Min. <fe Met. Eng. 
Trans., Vol 106, p. 99, 1933 



FUELS 105 

cines under the moving gas stream in the combustion zone and still 
have them spread uniformly over the hot bath of matte and slag. 
It was found that best results were obtained when the gun spout was 
inclined at an angle of 34 to 37 from the horizontal. The tempera- 
ture of the furnace was so great that it was impractical to use a sta- 
tionary feeder, and it was necessary to design a movable feeder which 
could be withdrawn after the charge had been dropped. 

The gun proper is made up of two sections, the upper part or 
carriage, and the water-cooled nose. These two parts telescope to- 
gether, with the carriage fitting inside the movable spout. The spout 
is moved in and out of the furnace by a rack and pinion drive. The 
gun carriage rides on tracks supported from abo\ e, and the calcine cars 
feed the calcine through the spout by means of a Tacoma dustless 
connection. Counterweights are used to allow easy movement of the 
gun in and out of the furnace. Water is fed through flexible hose to 
the coil surrounding the movable spout in order to keep it cool. When 
the gun is withdrawn, the opening in the wall is covered with a water- 
cooled damper or gate. 

Five guns are used to a furnace, two on one side and three on the 
other; the charging ports are staggered to allow even distribution of 
the charge. One of the features of this method of feeding is that the 
operator must know the condition of his furnace at all times and feed 
the material at the proper time and place. The guns enter through the 
side wall, and there are no drop holes in the arch. This permits 
the building of a stronger arch and contributes materially to the life 
of the furnace. Note that this furnace has a magnesite crucible and 
that the skewback plates are water cooled. 

The purpose of this development was to find a method of charging 
which would permit the continued use of deep-bath smelting and would 
eliminate the disadvantages caused by the method of center charging 
through the arch (as illustrated in Fig. 5) . The gun feed introduces 
the hot calcine under the gas stream and spreads it over the bath with 
a minimum of dusting. The decrease in dusting, and the strengthening 
of the roof by the use of a ribbed arch and elimination of the drop 
holes resulted in much longer furnace campaigns. The average life 
of the center-feed furnace was 80 to 100 days, but with the gun-feed 
furnaces a campaign will last over 200 days. 

Fuels. The three principal fuels used in copper matting reverbera- 
tories are pulverized coal, fuel oil, and natural gas, as we have already 
noted. These fuels are burned by means of special burners, which 
blow the fuel into the hot combustion chamber. All three fuels must be 
thoroughly and intimately mixed with combustion air so that rapid and 
efficient combustion is possible. Rapid and complete mixing of air and 



106 SMELTING 

fuel permits the fuel to be burned with nearly the theoretical amount 
of air, and this means that the flame will attain the maximum tempera- 
ture; also such a flame will be short, and the maximum amount of 
heat will be liberated close to the burners where most of the unsmelted 
charge is found. As a rule, high-pressure primary air is introduced 
into the burner to disperse the fuel; the primary air is not sufficient 
for complete combustion, and secondary air is drawn in around the 
mouth of the burner to provide enough total air for combustion. Later 
we shall describe burners used for each of these fuels. 
The important facts about fuels may be listed as follows: 

1. Cost and availability of the fuel are often the principal factors 
in choosing the fuel to be used. 

2. The calorific power of the juel, or the amount of heat evolved 
when a unit weight or volume'of the fuel is burned. 

3. The amount and nature of the ash in coal. 

4. The calorific intensity of the fuel, or the temperature attained 
when the fuel is burned. 

5. The amount of air required for combustion. 

6. The amount and composition of the gaseous products of com- 
bustion. 

For copper smelting, of course, it is necessary that the fuel selected 
should be as cheap as possible and also that a continuous supply 
should be available so that operations may not be interrupted by 
temporary fuel shortages. When two or more fuels are available, 
comparison of cost is made primarily on calorific power, since in buying 
fuel the user is really purchasing heat units; but other factors such as 
convenience in handling and storing must also be considered. 

TABLE 1 
EQUIVALENTS OF VARIOUS HEAT UNITS 

1 Btu = 2520 Calorie 
= 252 calories 
= 2930 watt-hour 
= 5555 pound-calorie 
1 Calorie = 3.969 Btu 

= 1 163 watt-hours 
1 kilowatt-hour = 3411 Btu 

= 860 3 Calories 
1 horsepower-hour = 2545 Btu 

= 641 3 Calories 
1 boiler horsepower = 558 . Btu per minute 

= 140 7 Calories per minute 

(Btu per Ib) X 0.5555 = Calories per kg (solid and liquid fuels) 
(Btu per cu ft) X 8 90 = Calories per cubic meter (gaseous fuels) 
1 barrel of fuel oil (bbl) = 42 U. S. gallons 



FUELS 107 

Pulverized Coal. The coal should be pulverized to the finest 
practicable state of division, which ordinarily means about 90 to 95 per 
cent minus 100 mesh and 80 to 85 per cent minus 200 mesh. Coal 
cannot be stored for any length of time after pulverizing because the 
large amount of surface exposed to the atmosphere results in consider- 
able oxidation ; the heat evolved may cause the particles to agglomerate 
or sinter together, or if enough air is available, the coal may take 
fire. Coal is pulverized as it is used, and although part of it may 
be stored for a while in surge bins or feeder bins, it is never allowed 
to stand for much more than 24 hours before it is used. 

Almost any coal can be used for pulverizing; usually a good grade 
of bituminous coal is used for copper smelting. The coal is delivered 
in lump form, dried if necessary, crushed, and pulverized. The pul- 
verizing may be done in ball mills, hammer mills, or similar equipment. 
Usually the pulverizing mill is traversed by a current of air which 
sweeps the particles out when they have been ground sufficiently fine, 
and the pulverized coal may be transported to the burners either by 
the use of screw conveyors or by blowing the air-coal mixture through 
a pipe. A feeder bin is sometimes located above the burners and kept 
full of pulverized coal; coal is fed from the bottom of the bin directly 
into the burners. 

Figure 15 illustrates the pulverized coal burner used on the reverbera- 
tory furnace at Noranda. An air-coal mixture is fed into the burner 
through the inclined pulverized coal pipe and is struck by a horizontal 
blast of air from the " converter air " pipe. This air comes from the 
same blowers as the air used for the converters and enters the burner 
at a pressure of about 5 pounds per square inch About 35 per cent 
of the air required enters with the coal dust; this amount cannot be 
increased because the mixture would then be likely to explode. The 
converter air amounts to from 45 to 50 per cent of the total air, and 
the secondary air which enters below the burner amounts to about 15 or 
20 per cent of the total. The secondary air comes from a preheater 
through a hot air duct and enters the furnace at a temperature of about 
300 C. This burner mixes the coal and combustion air very thor- 
oughly and gives a short hot flame. 

Natural Gas. Natural gas makes a very good fuel for reverberatory 
smelting, but, of course, its use is limited to those localities which are 
served by pipe lines from natural gas fields. Natural gas consists 
largely of gaseous hydrocarbons; methane, CH 4 , is usually the prin- 
cipal constituent together with ethane, C 2 H (J , propane, C 3 H 8 , and 
ethylene, C 2 H 4 . The heavier hydrocarbons found in some natural 
gas are usually removed by condensation to form casing-head gasoline. 



108 



SMELTING 



In addition to the hydrocarbons, natural gas contains small amounts 
of carbon dioxide, carbon monoxide, oxygen, and nitrogen. Natural 
gas will contain from 700 tc 1400 Btu per cubic foot and gives a 
high flame temperature. 



Hot Air ! 
Duct I 






C.L Furnace 




' 1 

NTT/^ 


."i. 1 - - . 








j)#^ 


Jrfe 


M 








5r 


^ 




Pulverlz 
Pip 


_-_-^^ 


~ r ~i~ 



















Plan 



Converter 
Air Pipe 




Coal Pipe 




Burner 



General Arrangement 

(Bogga and Anderson, Can Mm Jour , p 191, April 1984) 

Fia. 15. Pulverized Coal Burner, Noranda Mines, Ltd. 

Figure 16 is a section of the inspirator burner used for burning 
natural gas in the reverberatones at Anaconda, Montana. These 
operate exactly like Bunsen burners. Gas enters the burner from the 
gas manifold through a 2-inch nozzle at 20 pounds pressure and the 
injector effect of the gas stream draws primary air into the burner 
tube where the gas and primary air become mixed. As the mixture 
of gas and primary air issues from the mouth of the burner into the 
combustion chamber it meets the stream of secondary air drawn in 
around the burner mouth. The amount of both primary and secondary 
air can be controlled by means of shutters. (The shutters for regu- 
lating the secondary air can be seen in Figure 17.) There are five 
burners to each furnace, and each burner has a maximum capacity of 
20,600 cubic feet per hour. 

Fuel Oil. Fuel oil is a liquid solution consisting primarily of hydro- 
carbons, or of sulfur, nitrogen, or oxygen derivatives of the hydrocar- 
bons. The fuel oil may be either crude petroleum or a product of the 
refining of petroleum; sometimes the lighter fractions, such as gasoline, 



FUELS 



109 



or the heavy fractions, such as grease, wax, or asphalt, are removed 

from the crude oil, and the " middle 

fraction " used as fuel oil. Fuel 

oil is a very good fuel, burns clearly, 

and gives a high flame temperature. 

The heating value varies somewhat, 

and will average around 6,000,000 

Btu per barrel of 42 gallons. 

As in the case of powdered coal, 
it is essential that the fuel oil be 
intimately mixed with the combus- 
tion air if the burning is to be rapid 
and efficient; to attain this end the 
oil must be atomized and the mix- 
ture of air and finely dispersed oil 
droplets sprayed into the combus- 
tion chamber. Viscous oils may 
require heating before entering the 
burners so that the fluidity may be 
increased enough to permit efficient 
atomization. The atomization of 
the fuel is an important factor in 
the combustion of oil and is es- 
sentially parallel to the pulverizing 
of coal. High pressure air or steam 
jets are used in some cases; in other 
cases the bulk of the atomization is 
performed by mechanical methods. 
For copper smelting a low-pressure 
burner is generally used; i e. a 
burner which does not require high 
pressure air or steam to atomize the 
oil. The oil is heated to a tempera- 
ture of 120 to 150 C and delivered 
through pipes to the burner at a 
pressure of about 50 pounds per 
square inch. When this oil issues 
from the pipe into the burner it 
has a low viscosity and high vapor 
pressure and is readily dispersed by 
means of low-pressure primary air 
(a few nounds per square inch). The dispersion of oil and primary 




I luaujao /jopjjjiy I 

aflin MJI* i?s P" wr -g -I 
*oi *1 $2 

<* 

I! 

4 




110 



SMELTING 



air enters the combustion chamber, where it meets the secondary air 
drawn in around the mouth of the burner. 

Burners of all types are set in the end wall of the furnace, as shown 
by the various figures. At the Flin Flon smelter in Manitoba, auxiliary 
burners have been set in the side walls of the furnace near the back. 




(Courtesy Anaconda Copper Mining Company) 

FIG. 17. Natural Gas Burners on Anaconda Reverberatory Furnace. 

One of these is set on each side of the furnace about 30 feet from the 
bridge wall and at an angle of 60 from the long axis of the furnace. 14 
These have proved to be a valuable aid to the end burners both in in- 
creasing the tonnage treated and in smelting " floaters " on the slag bath. 

It is difficult to give definite figures on the amount of fuel required 
to smelt a ton of charge, as this depends largely upon the nature of the 
charge itself. Table 2 gives the data on fuel consumption at six 
typical plants. Attention is called to a few important points illus- 
trated by this tabulation. 

1. The amount of fuel used in a reverberatory furnace is commonly 
expressed as the juel ratio, or the amount of fuel (tons of coal, barrels 

14 Boggs, W. B., Copper Metallurgy Reveals Improvements: Eng. and Min. Jour., 
Vol. 139, No. 2, p. 70, 1938. 



FUELS 111 

of oil, or cubic feet of gas) required to smelt a ton of dry charge; the 
fuel ratios can be calculated in each case from the data in Table 2. 
Sometimes the inverse ratio, or the number of tons of dry charge 
smelted per unit of fuel, is used. The earliest furnaces required about 
0.50 ton of coal to smelt 1 ton of charge; modern reverberatories smelt- 
ing hot calcines will consume as little as 0.12 ton of coal per ton 
of charge. 

2. It will be noted that wet-charge smelting requires more heat 
units per ton of ore smelted than dry-charge smelting. However, 
only a part of the heat supplied is actually used in the furnace; part of 
it is used by the waste-heat boilers which are heated by the waste gases 
from the furnace. When wet charging is practiced there is a greater 
amount of heat abstracted by the boilers, and the net heats actually 
used for smelting will not differ as much as the gross heats. Table 3 
on page 114 will illustrate this point. 

Note that there is a much greater waste-heat steam recovery in the 
wet-charged furnace per ton of charge smelted. In this particular 
case, when the fuel used in drying is included, the net heat required 
for smelting 1 ton of charge is actually less in the wet-charge smelting 
than in the other two methods. 

3. Two examples in Table 2 deal with furnaces at Anaconda at the 
time the change was made from coal to gas firing. Note that there 
is very little difference in the number of Btu required per ton of charge 
smelted. The natural gas appears to be slightly more efficient than 
coal when comparing the gross or high calorific powers of the two fuels, 
and the difference is even greater when the low or net calorific powers 
are compared. 

4. Table 2 illustrates the effect of the charge on the gross fuel con- 
sumption of the furnaces. At Noranda, for example, 2,430,000 Btu 
is required per ton of charge; the charge is 93 per cent hot calcine and 
the secondary air is preheated. On the other hand, at Miami the 
charge is 81 per cent wet concentrate and cement copper, and here 
5,600,000 Btu is required per ton of charge. 

The temperature in reverberatory furnaces is usually about 1400 
to 1700 C in the smelting zone or firing end and 1100 to 1300 C at 
the flue end. The high temperature of the gases entering the flue means 
that a good deal of the heat evolved in combustion is not utilized in the 
furnace itself; and it is standard practice to pass the flue gases through 
waste-heat boilers and convert this heat into steam which can be used 
for various purposes around the plant. 

The distribution of the heat evolved in a reverberatory furnace is 
important; this is usually shown by means of a heat balance such as 



112 



SMELTING 



c, I 



co a 

- 

&< 

o 

fc 

o 



53 



2.1 j 



! d 2 

6 






II? 

U PH 



c >> 

1 



11 






- 

PQ 



-- -- 

PQ PQ 



1 



- 

O 

3 



c 



o 

I 

CO 
CO 



8 
1 



rge 
; 



-a S 



wet concentrate; 
4 0% is cement 



<M ' ^ 00 ^ 

(N o '~ H > - 1 o ' ' o 

T-H iO '^ i I LQ " 






qj X ^2 X3 

ft co c w o r>- 

s s 






$ , 



i-il 



o 

"CJ 



"8 



1 



1 



i 






FUELS 



113 



1-2 
1 1 1 1 



ih $ 

311 



n 03 v 

C J3 " 



BS 

6 



o; b 

^ 





S 



1 



ffl 




Pulverized 
coal. 



. 

l 



8 8 



O eo 

I" 1 



w "! 

8 -3 



I S 

w 



H S 

W | 
" 

s s 

2 -g 

S 8 

~ I 

a ? 

I I 

0) a> 

i I 



i . . a 

i en 3 

i^2 ( 



' ^ g 

5 ! 






114 



SMELTING 



TABLE 3 

FUEL CONSUMPTION WITH THREE METHODS OF CHARGING AT 
MIAMI, ARIZONA (Oil Firing) 





1931; 6 Month 


1924; One 


1926; One 




Period; 


Year Period; 


Year Period; 




Wet Charge 


Dry Charge 


Partial Roast 


Gross fuel ratio per ton dry solid 








charge at reverberatories 


902 bbl 


667 bbl 


0.577 bbl 


Equivalent waste-heat steam 








recovery 


0.470 


0.263 


219 


Net fuel ratio per ton dry solid 








charge at reverberatories 


432 


404 


358 


Drier fuel consumption 




105 


077 


Net total plant fuel ratio per dry 








ton smelted 


432 


509 


0.435 



Honeyman, P D I , op cit 

that given in Table 4. Of course, this distribution will be different 
for various furnaces and will depend upon such factors as the nature of 
the charge, flux, and fuel; whether the charge is wet concentrate, dried 
concentrate, or hot calcine; temperature and composition (hence total 
heat content) of matte and slag; temperature of the flue gases; nature of 
refractories used in walls and arch ; and whether combustion air is pre- 
heated or not. Table 4 is a heat balance for a coal-fired reverberatory 

TABLE 4 

HEAT BALANCE ON AVERAGE RESULTS, REVERBERATORY 
SMELTING, 6 MONTHS ENDING JUNE 30, 1920 





Btu per 
Furnace Day 


Per 
Cent 


Input: 






Coal 


1,827,676,000 


94 63 


Heat in calcine 


103,680,000 


5 37 


Total 


1,931,356,000 


100.00 


Output: 






Slag 


335,820,000 


17.40 


Matte 


96,000,000 


4.97 


Limestone (decomposition) 


96,567,800 


4.99 


Boilers 


579,406,800 


30.00 


Waste gases 


347,644,100 


17 94 


Radiation, conduction, and other losses 


475,917,300 


24.70 


Total 


1,931,356,000 


100.00 



Laist, Frederick, op. cit., p. 73. 



THE PRODUCTS OF THE REVERBERATORY FURNACE 115 

smelting hot calcine. Note that about 48 per cent of the total heat 
passes out in the gases and that 30 per cent is absorbed by the boilers ; 
in wet-charge smelting this figure would be even greater. 

The Products of the Reverberatory Furnace. Three products are 
removed from a reverberatory smelting furnace matte, slag, and 
fiuegas. The matte is of course the most important economic product, 
butTlEhe slag and flue gases are probably of greater importance in 
the furnace and plant operation. We shall take up the subject of 
matte and slag presently; after taking up converting and fire re- 
fining we shall devote a separate chapter to the consideration of flue 
gases. 

MATTE. We have already had occasion to refer to copper matte 
and have indicated that it consists essentially of artificial sulfides of 
copper and iron. The two sulfides which are stable at furnace tempera- 
tures are FeS and Cu 2 S. In Chapter II we assumed that these were 
the only substances present and made our calculations accordingly. 
These two liquid sulfides are soluble in one another in all proportions. 

The matte produced in copper smelting may range in grade from 
about 20 to 80 per cent copper. The solidified matte resembles in 
color and luster the massive form of the natural sulfide minerals. Low- 
grade mattes (20 to 50 per cent copper) show a dull bronze color on a 
fresh fracture, mattes containing about 60 per cent copper have a 
bluish-purple color, and mattes containing more than 70 per cent 
copper are almost white. High-grade matte which approaches the 
composition of Cu 2 S is called white metal. Many high-grade mattes 
also contain visible stringers of metallic copper or " moss copper." 
Mattes range in specific gravity from about 4 8 to 5 6, the gravity in- 
creasing with the copper content. Molten matte will contain as total 
sensible heat from 350 to 400 Btu per pound; this of course will de- 
pend upon the temperature and composition of the matte. A figure 
of 375 Btu per pound was used in computing the balance shown in 
Table 4. Pure Cu 2 S melts at 1100 C, and FeS at 1193 C; other 
mattes melt at lower temperatures, and according to the diagram of 
Carpenter and Hayward (Fig. 18), mattes containing from 30 to 50 
per cent Cu 2 S melt at about 1000 C. 

Figure 18 is an equilibrium diagram between the compounds Cu 2 S 
and FeS. It appears from the figure that FeS is quite soluble in Cu 2 S 
in the solid state, that Cu 2 S is less soluble in solid FeS, and that a eutec- 
tic of the two solid solutions exists. This diagram does not indicate 
the presence of any other compounds. Other in /estigators have pub- 
lished results which do not entirely agree with the deductions which 
would be drawn from Carpenter and Hayward's diagram (Fig. 18). 



116 



SMELTING 



Gibb and Philp 15 reported the presence of the stable compound 
5Cu 2 S'FeS in mattes, and in a more recent paper Avetisian 16 reports 
that the only stable compound formed is 2Cu 2 S*FeS. It would seem 
that there is as yet no complete analysis of the system Cu 2 S-FeS. 



1200 
1100 

fiooo 

c 

O 

o 
900 

3 
(0 

800 

O) 
H- 

700 
600 

BS* 10 








\ 


s. 














lf ^~~~* 




\ 

\ 




k 


095 Deg 


ees C. 




^^ : 


"^--' 








Cu,SlpFeS 
Plus Eutectlc 


ffjjS 


nCu z S 
E_u tectic 


Dimoi 


phic Char 


ge in Cry 


itallizatior 


of Cu 2 S 


1 


TransformationI 950 D 
I ~,i_ 


igrees C 










jij 


Cu 2 S u 
Plus Eu 


3\s 
FeS s ' FeS 
tectic JS \& Plus 


in Cu 2 S 
Eutectic 


Solid Sol 


ution FeS 


in Cu 2 S 




JSJ 




11 












1 

1 




lil 

sis' 












90 80 70 60 50 40 30 20 10 
10 20 30 40 50 60 70 80 90 100* Cu 2 S 



Per Cent Composition by Weight 

(Reproduced by permission from H of man and Hayward, Metallurgy of Copper, p 166, 

McGraw-Hill Book Co , New York, 19%4) 

FIG. 18. The Cu 2 S-FeS Equilibrium Diagram. 

It is likely that even complete knowledge of the Cu 2 S-FeS binary 
system would be inadequate, and it would probably be necessary to 
have an analysis of the 3-component, Cu-Fe-S system for complete 
interpretation. Figure 18 is determined from melting point data, and 
the liquidus (upper) curve gives us the melting points of mattes of 
different compositions. 

Although our previous assumption that copper matte consists of 
Cu 2 S and FeS in varying proportions is sufficiently accurate for many 
purposes, it is not strictly true; matte is a more complex substance 
than this would indicate. We shall not have space to consider all the 
investigations which have been made on the composition of mattes but 
shall merely present a few important facts. 

1. As a rule the combined percentages of copper, iron, and sulfur 
in matte will equal or exceed 95 per cent. 

2. The sulfur content of matte is usually less than would be expected 
theoretically, by calculating the amount required to form Cu 2 S and 

15 Gibb, Allan, and Philp, R. C., The Constitution of Mattes Produced in Copper 
Smelting- Am. Inst. Min & Met. Eng Trans., Vol. 34, p. 665, 1906. 

16 Avetisian, C. K., Copper Matte Composition: Eng. and Min. Jour., Vol. 133, 
No. 12, p. 627, 1932, and Vol. 134, No. 1, p. 27, 1933. 



THE PRODUCTS OF THE REVERBERATORY FURNACE 117 

FeS with all the copper and iron present. One reason for this is that 
some of the iron may be present as magnetite (Fe 3 4 ) or copper 
ferrite (CuOFe 2 03). Another reason is that sulfur appears to vola- 
tilize from the FeS, leaving behind metallic iron which is soluble in 
the remaining FeS. This excess iron can react with Cu 2 S thus: 

Cu 2 S + Fe -> 2Cu + FeS 

and this reaction may account for the moss copper found in some 
mattes. 

3. Matte may contain up to 10 per cent magnetite. It is not clear 
whether this substance is actually soluble in molten matte or not. 
Magnetite has a specific gravity of 5.1, which means that its density 
lies between the maximum and minimum densities of mattes; it will 
settle out through low-copper mattes, but it will float on high-copper 
mattes and be removed with the slag. In any event there is not much 
difference in density, and especially when the matte has about the same 
density as magnetite (Cu 30 to 40 per cent) we should expect to find 
the magnetite mechanically entrained in the matte. 

4 Many mattes contain metals other than copper and iron in im- 
portant amounts. When there is zinc on the charge, part of the zinc 
will enter the matte as ZnS; some copper mattes contain from 2 to 5 
per cent zinc (Table 5). In the smelting of nickel-copper ores all the 
nickel enters the matte as a sulfide, and as an approximation we may 
consider that these mattes consist of Cu 2 S + FeS -f NiS, just as we 
considered copper matte to be Cu 2 S -f FeS. Sometimes the formula 
Ni 3 S 2 is given for the nickel sulfide, but there probably is no such 
compound, and this represents the approximate composition of a solu- 
tion of Ni in NiS. In lead smelters, the copper is usually collected in a 
matte, and these byproduct mattes will contain considerable lead as 
PbS and as metallic lead. 

5. The other elements found in matte may be considered impurities, 
as the percentages are usually small. These will be determined by the 
nature of the charge and may include small amounts of cobalt, nickel, 
arsenic, antimony, bismuth, and lead. Arsenic and antimony are only 
slightly soluble in matte and when present in any quantity these ele- 
ments form a speiss which is insoluble in matte and separates from it 
in a distinct layer. Speiss is primarily an artificial arsenide or anti- 
monide, just as matte is an artificial sulfide. The formation of speiss 
is uncommon in reverberatory copper smelting; in most cases the small 
amounts of arsenic and antimony on the charge are dissolved in the 
matte. Most mattes also contain small amounts of Si0 2 , CaO, and 
other slag-forming substances. 



118 SMELTING 

6. All commercial mattes are excellent solvents for the precious 
metals gold, silver, and the platinum group metals and all these 
metals on the charge will be efficiently collected by the matte. Prac- 
tically the only possibility of losing any precious metals in copper 
smelting is in producing such a small amount of matte that it does 
not have an opportunity to come in contact with all the precious metals 
and collect them; the operator always aims to produce a matte of low 
enough grade and large enough volume to collect all the precious 
metals. 

The amount of matte produced expressed as percentage of the 
total charge is known as the matte fall; the per cent of copper (or 
copper + nickel for copper-nickel mattes) is the grade of the matte. 
For some of the mattes shown in Table 5 the theoretical amounts of 
iron and sulfur have been calculated according to the method of Chap- 
ter II, assuming that the matte consists only of Cu 2 S + FeS (and 
ZnS in case g) ; the theoretical figures are given in parentheses. Note 
that the actual iron assay checks closely with the theoretical but that 
the chemical assay for sulfur is always lower than the theoretical. 
This is especially evident in the low-grade mattes b and g ; apparently 
this is because a large part of the iron is present either as an oxide or 
as metallic iron dissolved in the iron sulfide. 

SLAG. Table 5 also gives the composition of a number of slags 
formed in reverberatory smelting; in each case the corresponding slag 
and matte are from the same operation. Note that in general these 
slags contain from 30 to 38 per cent silica, 45 to 52 per cent FeO, 5 to 8 
per cent A1 2 3 , and 1 to 5 per cent CaO. Essentially these slags are 
molten solutions of ferrous silicates in which are dissolved smaller 
amounts of other basic oxides (A1 2 O 3 , CaO, and MgO). Slag from 
the Roan Antelope deviates quite markedly from the average com- 
position. 

Requirements of a Slag. We have already noted that one of the 
functions of the reverberatory furnace is to permit slag and matte to 
separate as completely as possible; the ideal slag is the one which pro- 
motes the cleanest separation. Such a slag must have the following 
characteristics: 

1. Low specific gravity. The slag must, of course, be lighter than 
the matte, and the greater the difference in density (viscosity remain- 
ing the same) the more rapid and complete will be the separation of 
the two liquids. Reverberatory slags will range in specific gravity 
from 2.8 to 3.8, the heavier slags being those which are highest in iron. 
For the common slags containing 45 to 50 per cent FeO the specific 
gravity will usually be 3.3 to 3.5. 



THE PRODUCTS OF THE REVERBERATORY FURNACE 



119 



TABLE 5 

ASSAYS OF SOME TYPICAL MATTES AND SLAGS FROM 
REVERBERATORY SMELTING 

Mattes 



Plant 


Per Cent 


Cu 


Fe 


S 


Zn 


SiO 2 


Miami. 


45 9 


26 


25 2 










(27 0) 


(27 0) 







Noranda. 6 


19.3 


48 3 


22.5 


. . . 


. . . 






(48 0) 


(32 5) 






Roan Antelope. 


78.73 


58 


19 5 




0.14 






(075) 


(20 5) 






Andes. d 


45 88 










McGill. 6 


31.54 


... 


.... 






Anaconda/ 


42 00 










FlinFlon.' 


22.00 


41 4 


24 5 


4.2 








(40 0) 


(32 8) 







Slags 





Cu 


SiO 2 


A1 2 3 


FeO 


CaO 


MgO 


S 


Silicate 
Degree 


Miami a 


0.52 


34 5 


7 7 


49 8 


1 26 






1 22 


Noranda. 6 


0.27 


37 9 


8 5 


46 7 


1 7 


1.9 


1.1 


1 29 


Roan Antelope. 


1.20 


47 3 


14 7 


13 7 


11.4 


2 9 




1.75 


Andes. d 


0.60 


30.9 


7 2 


50 9 


2 5 


j> . 


8 


1 07 


McGill. 6 




36 1 


5 2 


48 8 


5 7 






1 29 


Anaconda f 


50 


32 


8 


52 


1 5 






1 07 


Flin Flon.' 


36 


36 5 


5 1 


41 3 


1 7 


4 7 


4 


1 39 



Honey man, P D I , op cit , p 88 

6 Bogga, W B , and Anderson, J. N , op cit , p 165 

c Wraith, C R , op cit , p 202 

<* Callaway, L A , and Koepel, F N , Metallurgical Plant of the Andes Copper Mining Company: 
Am Inst Mm & Met Eng Trans , Vol 106, p 090, 1933. 

* Larson, Leonard, op cit 

^ Bender, L. V , Development of Copper Smelting at Anaconda: Eng. and Min. Jour , Vol. 128, 
No 8, p 301, 1929 

Ambrose, J H , The Flm Flon Copper Smelter- Canadian Inst Mm Metallurgy, Bull 281, 
p 418, September 1935 

2. Low melting point. The slag must be completely molten at the 
furnace temperature. In early practice the slag was skimmed through 
side doors in the furnace by means of rakes and rabbles; such a slag 
could be removed even if not completely molten, although, of course, 
these slags were not as clean (free from copper) as could be desired. 
In modern practice, however, the slag flows from the furnace, and it 
must be completely molten. 



120 



SMELTING 



3. Low viscosity. Slags must be sufficiently fluid to flow easily from 
the furnace and to permit entrained globules of matte to settle rapidly. 
A slag may have a low melting point and low density, but if the liquid 
is " thick " or viscous the separation of slag and matte will not 
be clean. 

4. Low solubility for matte and metal. Most reverberatory slags 
have practically no solvent power for liquid matte or for metallic 




-Liu. 19. Dumping Reverberatory Furnace Slag. 

copper, but there is some evidence that a very small amount of 
sulfide may be in true solution in silicate slags. Oxidized copper, 
however, dissolves readily in slags; the copper oxide forms either copper 
silicate or copper ferrite, and these compounds dissolve readily in 
other silicates. The loss of copper in slag is high in any smelting opera- 
tion where oxidized copper is present in the charge. 

Slag Composition. The various acid and basic oxides which make up 
slags form " alloys " with one another, and equilibrium diagrams can 
be constructed which indicate the melting points of slags of various 
compositions, the nature of the solid phases formed, etc; these diagrams 
are essentially the same as those which are so widely used in the study 



THE PRODUCTS OF THE REVERBERATORY FURNACE 121 

of metallic alloys. Practically all copper slags, however, contain at 
least four important components (FeO, Si0 2 , Cat), and A1 2 3 ), and 
the analysis of four-component systems by the method of equilibrium 
diagrams is difficult and tedious. There has never been a complete 
theoretical investigation of copper slags, and it is doubtful if such an 
investigation would be worth the effort. Years of experience have 
determined the approximate slag composition which gives the best 
results, and, fundamentally, the method of selecting a slag composition^ 
is to use one which has already worked successfully. 

Silicate Degree. Slags are often classified according to their silicate 
degree, which is defined as the ratio of the weight of oxygen in the acid 
oxides in the slag to the weight of oxygen in the basic oxides, i.e.: 

Weight of oxygen in acid 

Silicate degree = . 1 

Weight of oxygen in bases 

The following names are given to slags according to their silicate degree. 

SILICATE DEGREE SLAG NAME 

<1.0 Suhsiiicate 

1.0 Monosihcatc or singulosilicate 

1.5 Sesquisilicate 

2 Bisnhcate 

3 Trisihcate 

For example, the simple slag CaSi0 3 or CaOSiOo has two atoms of 
oxygen in the acid (Si0 2 ) and one atom of oxygen in the base (CaO). 
Hence the silicate degree is f = 2, and the slag is a bisilicate. The slag 
Ca 2 SiO 4 or 2CaOSi0 2 has a silicate degree of | = 1; this slag is a 
singulo-silicate. For all practical purposes, Si0 2 is the only acid 
radical in copper slags; A1 2 3 may act as either an acid or a base 
(forming respectively aluminum silicates or metal aluminates), but 
it behaves as an acid only in very basic slags. In all our copper slags 
we shall consider A1 2 3 as a basic oxide. Consider the three slags: 

COMPOSITION PER CENT SiO2 

CaO-SiO 2 51 8 

FeO-Si0 2 45 5 

(CaO, FeO) SiO 2 45.5 to 51.8 

All these are bisilicates, but the per cent of silica is not the same in 
each slag. In the (CaO, FeO)-Si0 2 slag the per cent of silica may 
range between 51.8 and 45.5, depending upon how much of the CaO 
is replaced by FeO. In other words, two slags having the same silicate 
degree need not have the same weight percentage of silica, but they do 



122 SMELTING 

have the same molecular ratio of Si0 2 to basic oxides. The fact that 
the amount of oxygen is used as a basis of calculation takes care of 
valence changes in the metal radical, since oxygen has a constant 
valence of two. Thus an aluminum bisilicate would have the formula : 

Al 2 Si 3 9 or Al 2 3 -3Si0 2 

Let us now calculate the silicate degree of the Noranda slag in Table 5 
for one more illustration. The analysis as given does not add to quite 
100 per cent because some of the minor elements are not reported. We 
shall calculate the silicate degree on the basis of the analysis of the 
five principal oxides as given. One pound of this slag will contain 

0.379 Ib of Si0 2 
0.085 Ib of A1 2 3 
0.467 Ib of FeO 
0.017 Ib of CaO 
0.019 Ib of MgO 

32 

The weight of oxygen in the acid = 0.379 X = 0.202 Ib. 

60 

The weight of oxygen in the bases is 

A1 2 O 3 0.085 X 7^ = 0.040 Ib 



FeO 0.467 X -- = 0.104 Ib 

71.8 

i r* 

CaO 0.017 X = 0.005 Ib 
56 

MgO 0.019 X-7 = 0-008 Ib 



Total 0.157 Ib 

Silicate degree = ' =1.29 
0.157 

The silicate degree indicates the relative acidity of a given slag, and 
the comparison of slags by means of the silicate degree is actually 
based on the assumption that equivalent amounts of one base may be 
substituted for another without affecting the properties of the slag. 
This, of course, is not true calcium silicate and ferrous silicate are 
quite different substances but within the rather narrow limits of 



THE PRODUCTS OF THE REVERBERATORY FURNACE 123 

composition found in copper slags, the silicate degree is a useful 
criterion. 

Most copper slags approach the composition of a sesquisilicate (sil- 
icate degree = 1.5) as an average. The silicate degree may range 
from 1.0 to 2.0 but seldom passes either limit. The calculated silicate 
degrees are listed in Table 5; for these slags the values lie between 1.07 
and 1.75. 

Some Properties of Copper Slags. Copper smelting slags are black 
and have either a stony or a glassy appearance; slags which are rapidly 
cooled are glassy. As we have already noted, these slags are fairly 
heavy (specific gravity 2.8 to 3.8). 

Copper slags are usually discharged from the furnace at 1100 to 
1300 C, and the slag must be molten and free-running at this tempera- 
ture. The exact temperature of the melting point of different slags 
is difficult to obtain. The formation temperature of a given slag is 
probably of more importance than its actual melting point; this 
formation temperature is the temperature at which molten slag will 
form from the mechanical mixture of solid oxides on the charge. It 
is always higher (usually 100 to 300 C) than the melting point de- 
termined on a sample of the formed slag, and the coarser the pieces 
of slag-forming oxides the higher the formation temperature will be. 
In an operating furnace, the actual mechanism of slag formation is 
essentially the dissolving of the oxides ("earthy" material) in the 
pool of slag which is always maintained. 

As a general rule the more acid slags have a greater viscosity. Basic 
slags are " thm " and fluid; acid slags are " thicker " or more viscous. 
Basic slags are more corrosive to furnace refractories than acid slags. 

Copper slags as discharged from the furnace will contain from 500 to 
600 Btu of sensible heat per pound of slag; for the heat balance in 
Table 4, a figure of 579 Btu per pound was used. As is true of mattes, 
this value depends upon the temperature of discharge and the composi- 
tion of the slag. 

Copper Losses m Slag. As the slag from the smelting furnace is 
discarded and any copper contained in it is lost, the question of the 
manner in which copper is carried in the slag is of great importance. 
There are three principal ways in which copper can be carried out in 
slag, viz.: 

1. Oxidized copper (oxides, carbonates) form copper silicates and 
ferrites which are soluble in the molten silicate slag. 

2. Silicate fusions appear to have a slight but definite solvent power 
for sulfides, so that some of the matte may be actually dissolved in 
the slag. 



124 SMELTING 

3. Small particles of matte may be mechanically entrained in the 
slag and swept out of the furnace with the slag stream before they have 
an opportunity to settle. 

In a recent paper on this subject, Jackman and Hay ward 17 report 
that most of the copper found in reverberatory slags is in the form 
of small irregular pellets of sulfides; metallic copper may be present at 
times, but it is of minor importance. They found no proof of the 
presence of copper silicate in the slags studied. 

It appears that when there is little or no oxidized copper on the 
charge, the copper lost in the slag is largely in the form of sulfide 
pellets. These can be identified in the solidified slag, but it may be 
noted that this fact does not indicate whether they were dissolved or 
mechanically entrained, because it is likely that sulfides which were 
soluble in the liquid slag would separate out when the slag solidified. 
Probably both factors contribute to the presence of sulfides in slag. 

The presence of oxidized copper on the charge usually leads to 
higher slag losses, because unless this copper can be either sulfidized or 
reduced to metallic copper it w r ill dissolve in the slag. At Roan Ante- 
lope (see Table 5) the concentrate carried 2.40 per cent Cu 2 0, and 
the original slag from the smelting contained 2.25 per cent copper, 
largely because of the slagging of the copper oxide. Later 3.0 per 
cent by weight of fine coal was added to the concentrate before 
charging; this served to reduce the oxide to metallic copper and 
lowered the slag assay to 1.20 per cent copper. 18 

As a general thing, the copper content of reverberatory slags will 
range from 020 to 0.60 per cent; in a few slags it will go above 1.0 
per cent. The copper assay of slag is roughly proportional to the per 
cent of copper in the matte the richer the matte the higher will be 
the copper content of the slag. This would be expected from the 
discussion given above. To illustrate this point the data in Table 5 
have been plotted in Figure 20. The copper content of the Roan 
Antelope slag is higher than might normally be expected even with 
such high-grade matte; two contributing reasons for this are (1) the 
presence of oxidized copper in the concentrate, and (2) the viscosity 
of the high-alumina slag produced. 

The total copper loss in slag depends upon both the copper content 
of the slag and the amount of slag produced (slag volume), and of 
course both are important. Roan Antelope, for instance, finds it more 
profitable to make a slag of 1.20 per cent copper (Table 5) than to 

17 Jackman, R. B., and Hayward, C. R., Forms of Copper Found in Reverbera- 
tory Slags: Am. Inst. Min. & Met. Eng. Trans., Vol. 106, p. Ill, 1933. 

18 Wraith, C. R., op. cit., p. 215. 



THE PRODUCTS OF THE REVERBERATORY FURNACE 125 

lower the copper content by changing the slag composition. The use 
of more limestone flux would make the slag less viscous and would 
decrease its copper assay, but the volume of slag would increase. 
Wraith 19 states that in this instance the amount of copper lost in the 
viscous slag of high copper content and low volume is less than is 
lost in a fluid slag of low copper content and larger volume. 



1.2 



i.o 



i 

550.6 



>0.4 



0.2 



CO 




10 20 30 40 50 60 70 80 

Grade of Matte per cent Copper 

Fia. 20. Variation of the Copper Assay of Slags with the Grade of the Matte. 

Letters refer to the plants in Table 5 

Fluxes. In order that the slag shall have the desired composition 
it is often necessary to add flux to the furnace charge; sometimes the 
flux is charged directly into the reverberatory, but when the material 
is roasted the flux may be added at the roasters. The most satisfactory 
way of fluxing a reverberatory charge is to mix ores and concentrates 
of different composition so that the gangue minerals are present in 
the correct proportion to form a suitable slag. Thus Noranda 20 uses 
a siliceous copper ore to flux the high-iron ores and calcines; this 
combination produces a suitable slag (Table 5) without the use of 
any other flux. Very often, however, it is necessary to use a barren 
fiux, i.e., a material which carries no copper. The disadvantages of 

19 Wraith, C. R., op. cit., p. 215. 

20 Boggs, W. B., and Anderson, J. N., op. cit,, p. 165. 



126 SMELTING 

this are obvious, because every ton of barren flux charged displaces a 
ton of copper-bearing material and therefore decreases the smelting 
capacity for calcine or concentrate and increases the slag volume. Be 
that as it may, in most cases it is necessary to add a certain amount of 
barren flux to obtain a proper slag composition. The most common 
flux used is limestone, which contributes CaO to the slag. The other 
important oxides (FeO, Si0 2 , and A1 2 3 ) are usually present in the 
smelting charge. In modern plants there is seldom a deficiency of FeO, 
and when Si0 2 is needed it is usually supplied as siliceous ore; A1 2 3 
is never added intentionally, as it makes the slag viscous, and is 
generally an undesirable constituent of slags. 

Slag Disposal. The liquid slag tapped from the furnace may be 
either (1) collected in slag pots which are then hauled to the slag dump 
where the slag is poured or (2) allowed to flow directly into a large 
volume of water in a launder; the slag is granulated and sluiced to the 
dump. Slag may be used occasionally for such purposes as furnace 
foundations, but practically all slag is sent directly to the dump and 
discarded. Some of the slags produced by the smelting of oxide ore 
in the southwestern United States in the early days (1881-1890) 
contained enough copper (2.5 to 4 5 per cent) so that it later became 
possible to resmelt them to recover the contained copper. Present-day 
copper slags are probably too lean to be resmelted, but when ore sup- 
plies become exhausted it may be necessary to find some way to 
recover the copper in these slag dumps possibly by leaching or con- 
centration, if the copper is in such a form as to make these methods 
feasible. 

CHARGE CALCULATIONS. We shall consider a simplified example to 
illustrate the method used in proportioning the various components of 
the reverberatory furnace charge so that when the charge melts, slag 
and matte of the desired composition will be formed. We shall con- 
fine our illustration to one particular type of calcine, namely a high- 
iron, low-copper material which will be smelted to a low-grade matte. 
The smelting of high-grade calcine, or wet concentrate, would require 
a different type of charge, but the method of calculation would be 
the same. 

Materials to be charged to the furnace are usually stored in bins 
from which weighed amounts can be drawn as desired; the exception 
to this is hot calcine, which usually comes directly from the roasters. 
Proper amounts of each material are then either transported to the 
furnace in charging cars or else fed onto the conveyor belt system 
which serves the furnace. In some plants the charge is bedded, i.e., 
a large pile of material is built up containing layers of the different 



THE PRODUCTS OF THE REVERBERATORY FURNACE 127 

constituents of the charge in their proper proportions; this material 
is then excavated in such a way that those proportions are maintained 
and is charged into the furnace. 

EXAMPLE 1 

Let us assume that we have available a low-copper, high-iron calcine of the follow- 
ing composition: 

Per Cent 
Cu 80 

S 12.7 

SiO 2 17.0 

A1 2 O 3 3.0 

Fe 46.7 

Also there is available an ore containing quartz in large quantity, pyrite, and chal- 
copynte. The composition of this ore is as follows: 

Per Cent 
Cu 1.0 

S 17.2 

SiO 2 65.0 

A1 2 O 3 2.0 

Fe 14.8 

Pure limestone can be secured for fluxing purposes if needed Of the sulfur in the 
calcine we can expect a 10 per cent loss by volatilization, and 50 per cent of the sulfur 
in the siliceous ore will be volatilized. Let us neglect dust losses and copper losses 
in the slag; we shall also assume that the matte will consist of FeS and Cu2S. 

Let us first calculate the relative amounts of the two materials we should use to 
secure a 20 per cent matte. 
Assume 

100 Ib of calcine 
x Ib of siliceous ore 

Then the total copper will be 8.0 -f 0.010.C Ib and total sulfur to the matte will be 

0.9(12.7) + 05(0.172z) = 11.4 + 0.086.C 
A 20 per cent matte will contain (Chapter II) 

Per Cent 
Cu 20.0 

S 32.3 

Fe 47.7 

Therefore 

8.0 + 0.010s _ 20.0 
11.4 + 0.086J ~ 32.3 
12.9 + 0.01615z = 11.4 + 0.086x 
x = 21.5 



128 SMELTING 

Hence the charge should consist of 21.5 Ib of the siliceous ore for every 100 Ib of 
calcine. 

Let us see what sort of slag analysis this charge would give 

Total weight of copper on charge = 8 -f 0.010(21 5) = 8.215 Ib 
8.215 20 



y 47.7 
y = 19.6 



where y = no. of Ib of iron to matte 



Total iron to slag = 46.7 -f (0.148 X 21.5) - 19.6 = 30.3 Ib and the slag will be 
made up as follows: 

FeO = 30 3 X B = 39 Ib = 53.1% 

Si0 2 = 17.0 + (0.65 X 21.5) = 31.0 Ib = 42.2% 

A1 2 O 3 = 3.0 + (0.02 X 21.5) = 3.4 Ib = 4 6% 



73.4 Ib 99.9% 



The weight of the matte would be : 

8125 40.63 Ib 



0.20 

and the matte fall 

4063 



121 5 



X 100 = 33.5% 



If we have only these two substances to smelt, and if it is desired to have a 20 per 
cent matte, then only one combination can be used, namely the one arrived at pre- 
viously. However, this combination yields a rather acid slag Let us calculate the 
silicate degree. 

O 2 in SiO 2 = 422 X f = 225 Ib 

O 2 in FeO = 0.531 X || = 0.118 Ib 

O 2 in A1 2 O 3 = 0.046 X -fifa = 0.022 Ib 

Total O 2 in bases = 0.118 + 0.022 = 0.140 

0.225 

Silicate degree = - = 1 61 
0.140 

This slag might be too acid; this could be remedied by (1 ) adding a little limestone 
flux, or (2) cutting down the amount of siliceous ore used. The latter remedy would 
affect the grade of the matte somewhat, but that would probably be more desirable 
than to add a barren flux. The Al 2 Os content is low enough so that it is probably 
unnecessary to add lime; if, however, there were more A1 2 O 3 on the charge, the ad- 
dition of some lime would probably be called for. Let us cut the amount of siliceous 
ore to 18 Ib and see how this affects the grade of matte and the slag analysis. 

The total copper on the charge will now be 

8.0 4-0.01(18) = 8.18 Ib 

8.18 X -&fa = 2.04 Ib of 8 as CujR in matte 

Total weight of sulfur to matte will be 

11.4 + (0.086 X 18) = 11.4 + 1.55 12.95 Ib 

12.95 - 2.04 = 10.91 Ib of 8 as FeS in matte 
10.91 X ff = 30.0 Ib of FeS in matte 



THE PRODUCTS OF THE REVERBERATORY FURNACE 129 

Total weight of matte = 30.0 + 8.18 + 2.04 = 40 22 Ib 

8.18 



40.22 



X 100 = 20.3%, grade of matte 



40 22 

.-- X 100 = 34.1% = matte fall 
118 

The change in the amount of fluxing ore alters the matte fall and grade of matte 
only slightly Let us see what happens to the slag 

Total iron on charge = 46.7 + (0.148 X 18) = 49.36 Ib 
Iron to matte = 10 91 X |f = 19 1 Ib 

49.36 - 19 1 = 30.26 Ib of Fe to slag 

FeO = 30.26 X if = 39 Ib = 54.8% 

SiO 2 = 17.0 + (0 65 X 18) = 28 7 Ib = 40 4% 
= 30 + (0.02 X 18) = 34 lb = 4.8% 



71.1 Ib 100.0% 
The silicate degree of the new slag would be 

() 2 in 8i() 2 = 404 X f f = 0.215 Ib 

O 2 in FeO = 548 X |f = 122 Ib 

O 2 in A1 2 O 3 = 048 X T V 8 ? = 023 Ib 

Total O 2 in bases = 122 + 0.023 = 0.145 

0215 
Silicate degree = ------ = 1 48 

0.145 

This is probably a more suitable slag, and it might be advisable to make the slag 
still more basic and bring the silicate degree down to 1.2 or 1.3, it is evident from 
the calculations made so far that the amount of the fluxing ore could be decreased 
considerably without affecting the matte grade very much 

Finally, let us determine what the grade of the matte would be if we were to smelt 
the calcine alone, i e., use an acid flux of pure silica, so that no copper or sulfur would 
be introduced from the flux. 

Weight of Cu = 8.0 Ib 

Weight of S in Cu 2 S = 8.0 X tW = 2 - lb 

Total H to rnatte = 11.4 lb 

S to FeS - 11 4 - 2 - 94 lb 

Weight of FeS = 9.4 X || - 25 8 11) 

Total weight of matte = 25.8 + 8.0 + 2.0 = 35.8 lb 

o n 

-f- X 100 = 22.3% - grade of matte 
o5.o 

Hence the maximum grade of matte possible would be 22.3 per cent Cu; if the 
volatilization loss of sulfur were 10 per cent, as we have .assumed, no higher-grade 
rnatte could be produced from this calcine. 

The calculations and assumptions made so far illustrate the method 
used in proportioning the charge to a reverberatory smelting furnace; 



130 SMELTING 

these have been intentionally simplified so that the principles involved 
might not be obscured by too much arithmetic. In an actual prob- 
lem, the calculations would be longer and more tedious because of a 
number of factors which we have omitted. Let us briefly consider 
some of these. 

1. An actual ore or calcine would show more constituents than we 
have indicated; these would include such substances as CaO, MgO, and 
MnO, which would go to the slag; possibly some zinc and lead which 
would be distributed between the matte and slag; small amounts of 
antimony and arsenic which would be partly volatilized and partly 
dissolved in the matte; and tin, cobalt, nickel, etc, and precious metals, 
which would go into the matte. Small amounts of basic oxides may be 
counted as FeO or CaO; thus MnO may be figured as if it were FeO, 
and as the molecular weights are almost the same (70.93 and 71.84), 
it may be assumed that these two bases replace one another pound for 
pound. The molecular weight of CaO is 56 against 40.3 for MgO, so 
that one pound of MgO is chemically equivalent to 56.0/40.3 = 1.39 
pounds of CaO. Thus a flux containing 30 per cent CaO and 10 per 
cent MgO could be considered as containing 30 -f (10 X 1.39) = 43.9 
per cent equivalent or summated lime, which is often written 



2. Not only would the analysis of each component of the charge be 
more detailed than we have indicated, but there would be other ma- 
terials charged besides calcines and flux. Possibly there might be 
two or more different calcines, some direct smelting ore, and cement 
copper precipitate. Certainly there would be a considerable tonnage 
of reverts of one kind or another. With buch a low-grade matte there 
would be a large tonnage of molten converter slag to be returned to the 
reverberatory for removal of its copper content. This converter slag 
would contain all the iron in the matte in the form of an iron silicate 
slag formed in the converters; the silica content of this slag would 
probably be lower than that desired for the reverberatory slag, so 
that it would be necessary to add more siliceous flux than that required 
for the iron in the calcines. In addition to the converter slag, flue 
dust, matte skulls from the ladles, custom ores, refinery products, etc, 
would have to be charged into the reverberatory from time to time. 
The effect of all these materials upon the matte and slag would have 
to be considered. 

3. If powdered coal were used for fuel, the coal ash would enter the 
charge and become part of the slag. Part of this ash would fall 
directly on the charge and part of it would pass into the flue to return 
eventually with the flue dust. 



THE PRODUCTS OF THE REVERBERATORY FURNACE 131 

4. With a low-grade charge such as this, the slag volume would be 
high, and the slag would have to be as low in copper as possible. 
Probably, in a case like this we would expect the slag to run not 
over 0.3 per cent copper. The fact that the matte is of low grade will 
insure a low-grade slag provided the slag is fluid enough ; this probably 
would be the factor which determines whether or not a lime flux would 
be used, for the lime would not be added unless the values saved by 
making a cleaner slag were greater than the cost of diluting the furnace 
charge with barren flux. Taking the figures from the first part of 
the example, we find that with a slag assaying 0.2 per cent Cu, we 
would lose 73.4 X 0002 = 15 pound of copper out of 8.125 pounds, 
or a loss of 1.84 per cent; with a slag containing 0.6 per cent Cu the 
loss would rise to 73.4 X 0.006 = 0.44 pound, or 5.4 per cent of the 
total copper on the charge. 

5. We have assumed that all the iron on the charge becomes FeO and 
is slagged as such. With such a high-iron charge, it is certain that 
there would be a good deal of magnetite entering the furnace from both 
the calcine and the converter slag. If this were not reduced to FeO 
by the sulfides on the charge, part of it would settle through the low- 
density matte and accumulate on the furnace bottom and the rest 
would be removed with the matte and slag. The amount and dis- 
position of the magnetite would be of great importance in the operation 
of the furnace. 

6. There are several ways 21 of calculating furnace charges, ranging 
from a precise and formal algebraic method to a simple cut-and-try 
procedure. The latter is probably most commonly used, as the prob- 
lem is usually one of making slight changes in existing charges rather 
than calculating an entirely new charge. 

7. Let us consider briefly how the reverberatory furnace practice 
is tied up with other operations. The calcine is the type that would 
result from the roasting of ore or concentrate high in pyrite, pyrrhotite, 
or both. Unless the ore minerals were unusually fine grained it should 
be possible to mill such an ore to give a much higher grade of con- 
centrate by rejecting a large part of the iron sulfides; the low copper 
content of this calcine suggests that (1) the ratio of concentration 
is probably low perhaps the milling is designed simply to reject 
gangue minerals and remove the bulk of the sulfides or (2) a con- 
siderable amount of heavy sulfide ore goes directly to the roasters 
without milling. This would indicate that copper is not the only metal 
to be recovered but that there is considerable gold associated with the 

21 Butts, Allison, Textbook of Metallurgical Problems, McGraw-Hill Book Co., 
New York, 1932. 



132 SMELTING 

iron sulfides. The collecting of the gold would require a reasonably 
high matte fall. 

In other words, this particular example illustrates the type of smelt- 
ing we should expect for a heavy sulfide copper-gold ore. If copper 
was the only valuable mineral we would expect a different method of 
treatment, either (1) the bulk of the iron sulfides would be rejected 
in the mill to give a richer feed to the smelter or (2) a method would 
be devised to recover the iron and sulfur in the pyrite as well as the 
copper. It would also be possible to make a high-copper concentrate 
from an ore of this type if it were feasible to recover the gold from 
the sulfides in the mill tailing for example by regrinding and 
cyamdation. 

With such a heavy fall of low-grade matte a relatively large con- 
verter installation would be required to provide sufficient capacity. 
There are three stages at which iron and sulfur can be rejected 
(1) in the mill, (2) in the roasters and reverberatory, and (3) in the 
converters, and the process should be adjusted so that the combination 
of these produce the desired results with the minimum total cost. Of 
course the practice should be adjusted so that all existing equipment 
(roasters, reverberatories, and converters) operates at full capacity. 

If the siliceous ore contained considerable gold, this would be 
another incentive to make a large amount of low-grade matte because 
this would consume more of the siliceous ore both in the reverberatory 
and in the converters. Thus the mining practice would have its effect 
on smelting, for the smelting practice might be altered as the supply 
of the siliceous ore increased or decreased. 

The roasting practice will depend upon a number of factors. It 
may be desirable under some conditions to secure a low-sulfur dead 
roast which yields a relatively cool calcine, or it may be preferable 
to remove less sulfur and produce a " hot " calcine (in which the 
sulfides are still burning) and thus introduce more heat into the 
reverberatory. The amount of iron oxides formed in roasting must 
also be considered, for if the calcine is too highly oxidized there may be 
too much magnetite formed in the reverberatory. 

Thus it may be seen that the operation of the reverberatory de- 
pends upon a number of factors roaster and converter capacity, 
milling methods, mining practice, metal prices, and fuel costs, to 
name only a few. Changes in any of these may bring about cor- 
responding changes in smelting practice. 

EXAMPLES OF PRACTICE. We have already cited figures on fuel 
consumption, slag and matte composition, and furnace construction 
at several plants. Before we leave the subject of reverberatory smelt- 



THE PRODUCTS OF THE REVERBERATOR Y FURNACE 133 

ing let us briefly recapitulate the essential features of the practice at 
two smelters which employ different smelting techniques. 

Fhn Flon. 22 There is one reverberatory furnace (Fig. 8) at the 
Flin Flon smelter in Manitoba. The furnace is 104 feet 3 inches by 
25 feet 6 inches, outside measurements, and 101 feet 3 inches by 21 
feet 6 inches inside the brickwork. The bottom is built of crushed 
silica, side walls of silica brick with a magnesite brick insert in the 
smelting zone, and the arch partly of silica brick (sprung arch) and 
partly of magnesite brick (Detrick suspended arch). The furnace has 
a hammerhead design, and the gases turn at right angles in both 
directions at the flue end into two 768-horsepower Stirling waste-heat 
boilers. 

The calcine and flue dust are fed by means of feed hoppers which 
discharge through pipes along the side walls near the firing end. 
Matte is tapped through Anaconda-type water jackets set in one 
side wall; a cast-copper plate is set in each water jacket, and in the 
center of this plate is a magnesite brick 5 inches square and 2 l / 2 inches 
thick. A 2-inch hole through this brick is the taphole. Three watej 
jackets are provided, but only the two nearest the burner wall are 
used. Slag is skimmed from the furnace through the front wall; it 
runs through a 12-foot launder into 225-cubic-foot cast-steel slag 
pots which are hauled to the slag dump. Converter slag is charged into 
the furnace through a launder set in the firing end above the burners. 

Pulverized coal is used as fuel, and there are four burners. The 
charge consists principally of calcine, plant reverts, and liquid converter 
slag. The charge requires a siliceous flux and most of this enters the 
circuit at the roasters and converters. The furnace charge is low in cop- 
per, and a low-grade matte is produced; the analyses and amounts of the 
various smelter products are given in Table 6. At the end of 1934 the 
furnace was handling 1040 tons of solid charge per furnace day. 

Roan Antelope. 23 The Roan Antelope smelter at Luanshya, 
Northern Rhodesia, has one reverberatory furnace which is 100 feet 
by 25 feet, inside dimensions. Construction is silica brick throughout. 
Side walls are 2 feet thick and 7 feet high. The arch is 18 inches 
thick for a distance of 65 feet from the burner wall and 15 inches thick 
over the remainder of the furnace. A Stirling waste-heat boiler is 
connected to the uptake of the furnace, and this boiler utilizes 52 
per cent of the heat value in the coal burned. 

22 Ambrose, J. H., The Fhn Flon Copper Smelter: Canadian Min. Met Bull. 281, 
p. 402, September 1935. 

23 Wraith, C. R , Smelting Operations at Roan Antelope Copper Mines, Limited: 
Am. Inst. Min. & Met. Eng. Trans,, Vol. 106, p. 202, 1933. 



134 



SMELTING 



TABLE 6 

FLIN FLON 

Typical Assays 





Cu 


Zn 


Pb 


Fe 


SiO 2 


A1 2 3 


MgO 


CaO 


S 


Calcine 


8 00 


4.1 


0.7 


31.0 


20.9 


2.3 


4.6 


1.4 


13.3 


Roaster flue 




















dust 


7.43 


4.2 




28.5 


18.3 










Roaster Cottrell 




















dust 


6.86 


4.2 




23.6 


20.0 










Converter slag 


1.50 






39.8 


30 


4 4 


1.0 


1.7 


. . . 


Flux 




... 


... 


3.0 


74.0 


12.5 


0.6 


3.5 


... 


Reverberatory 




















slag 


0.36 


4.0 


0.2 


32.1 


36.5 


5.1 


4.7 


1.7 


0.4 


Matte 


22.00 


4.2 


1.6 


41.4 










24.5 























Weight of Material Charged During 1934 (85 furnace days) 

Roaster products 252,453 tons 

Other pay material 6,490 

Recharged material 11,377 

Flux direct 66 



Total solid charge 
Liquid converter slag 

Total solid and liquid charge 

Coal burned 

Coal, per ton of solid charge 
Matte produced 

Matte fall, per cent of solid charge 
Slag produced 

Slag produced, per cent of total 
solid and liquid charge 



270,386 tons 
162,522 

432,908 tons 

32,019 tons 

118 per cent 
126,284 tons 

46.7 per cent 
251,469 tons 

58 7 per cent 



This is a wet-charged furnace, and the principal constituent of the 
feed is a high-grade flotation concentrate which is primarily a mix- 
ture of chalcocite (Cu 2 S) and a shale gangue. Limestone is used as 
flux. The charge is fed into the furnace through 6-inch pipes which 
extend vertically downward from the overhead hoppers into holes in 
the roof adjacent to the side walls. There are 12 charge pipes on each 
side spaced at 5-foot intervals. Slag is skimmed through an opening 
in the front wall of the furnace, and the sill of the skimming door is 
24 inches above the lower matte taphole, and the surface of the slag 
may be raised 6 to 12 inches by means of temporary clay-mud dams 
built across the skimming openings. Molten slag is laundered into 



THE PRODUCTS OF THE REVERBERATORY FURNACE 135 



200-cubic-foot cast-steel ladles in which it is hauled to the dump. 
There are two tapholes for matte, one directly above the other. The 
tapping block consists of a magnesite brick cast in a block of copper 
and is set in the side of the furnace 75 feet from the burner wall; 
two 3-inch holes are drilled through the brick at 9-inch centers. These 
holes are sealed with a clay dolly in the usual manner; they are 
usually opened by burning out the hole with an oxygen lance; some- 
times a tapping bar is used. In order to prevent the formation of 
heavy skulls of frozen matte in the ladles it is desirable to tap the 
matte as quickly as possible, and when the matte does not flow as 
rapidly as desired from one taphole, both are opened. As a rule a 
22-ton matte ladle can be filled in less than 10 minutes. 

TABLE 7 

ROAN ANTELOPE 

Typical Assays 





Cu 


Si0 2 


FeO 


CaO 


Al20 3 


MgO 


S 


Concentrate 


58 67 


13 4 


3 90 


10 


4 12 


70 


16 02 


Limestone 




0.6 


17 


52 30 


1 02 


2 04 




Matte reverts 


78 73 




0.58(Fe) 








19.52 


Reverberatory flue 
















dust 


31.22 


19.82 


4.59 


3.13 


11 37 


0.35 


5.7 


Converter flue dust 


73.29 


88 


69 


22 


1 32 




11.41 


Reverberatory slag 


1.20 


47.3 


13.7 


11 4 


14.7 


2 9 




Matte 


78.73 


14 


58 (Fe) 








19.52 



















Operating data, February, 1933 



Dry charge per furnace day 

Coal consumed per day 

Fuel, per cent of charge 

Flux, per cent of concentrate 

Flux, per cent of charge 

Slag, per cent of charge (excluding reverts) 

Slag, per cent of charge (including reverts) 

Silicate degree of slag 

Matte fall, per cent of charge 



340 
52.7 
15.5 

6.5 

5 09 
25.5 
21 22 

1.704 
70.0 per cent 



tons 

do. 

per cent 

do. 

do. 

do. 

do. 



The matte reverts amount to about 3.3 per cent of the weight of the charge. 
Flue dust is added at irregular intervals and is not considered part of the regular 
charge. 

The fuel used is pulverized coal and there are four burners. The 
analysis and relative amounts of material charged and produced are 
given in Table 7. It is interesting to contrast these with the figures 



136 SMELTING 

in Table 6. Roan Antelope slag differs from most other reverbera- 
tory slags, as we have noted (Table 5), in that it is higher in A1 2 3 
and SiO 2 and lower in iron; the copper assay of the slag is high, but 
the slag volume is low. The large amount of shale gangue accounts 
for the high AloOy content of the slag. 

The charge is high in copper, the matte is almost pure Cu 2 S, and 
the matte fall is very high. As there is practically no iron in the 
matte, there is no converter slag to return to the reverberatory. 

The two examples cited were chosen because they illustrate what 
might be called two extremes in reverberatory practice. One smelts 
a low-grade charge high in iron to produce a low-grade matte, and 
the other treats a high-grade charge and produces an exceptionally 
high grade matte. Practice at most other plants will generally lie 
somewhere between these limits, although there are smelters which 
produce matte as low in copper as the Flin Flon matte or lower. 

We have devoted the major part of this chapter to reverberatory 
matte smelting for the reason that this is by far the most prevalent 
type of copper smelting. We shall now consider briefly the other 
types of copper smelting that are used. 

MATTE SMELTING IN THE BLAST FURNACE 

Introduction. For many years the blast furnace had been widely 
used for matte smelting, but since 1910 it has gradually been dis- 
placed by the reverberatory. This change has been largely due to 
the fact that more and more copper has been produced in the form of 
finely divided concentrate, which can be more readily handled in the 
reverberatory furnace. At Anaconda, for example, both blast furnaces 
and reverberatories were used from the beginning, but as the milling 
methods improved, the superiority of the reverberatory became evi- 
dent; in January 1919 the blast furnaces were shut down and the 
plant dismantled. 

The Blast Furnace; General. The blast furnace is a shaft furnace 
and contains a vertical column of the charge to be smelted; as the 
charge is smelted down the liquids formed settle to the bottom, from 
where they can be removed; new material is charged at the top in 
quantity sufficient to keep the charge level relatively constant, and as 
material is fused at the bottom of the column, the column descends 
and more new material is charged on top. 

The fuel for the blast furnace is almost invariably coke, and it is 
charged from the top along with the ore and flux. In the side walls 
of the furnace are located the tuyeres through which the air or blast 
enters the furnace; the tuyeres are always located near the bottom of 



THE BLAST FURNACE GENERAL 137 

the furnace but are high enough above the bottom so that any liquids 
which collect in the furnace can never rise above the tuyere level. 

Blast furnaces have different shapes; iron blast furnaces are circular 
with the tuyeres spaced evenly around the circumference, whereas 
lead and copper blast furnaces generally have rectangular cross- 
sections with tuyeres along both sides but not on the ends. The walls 
of iron blast furnaces are constructed of refractory brick; non-ferrous 
blast furnaces usually have steel water jackets for the walls of the 
lower portion, and the upper walls may be constructed of refractory 
brick or they may consist of a second tier of water jackets. Blast 
furnaces may have an internal crucible directly beneath the tuyeres 
where the liquids collect and separate into layers, or the liquids may 
flow out of the furnace together directly beneath the tuyeres and pass 
into an external crucible or forehearth, where the separation takes 
place. Copper blast furnaces for matte smelting usually have ex- 
ternal crucibles. 

Before describing the copper blast furnaces, let us consider some 
of the general characteristics of blast furnace smelting. 

1. The hottest part of the blast furnace is the region just above 
the tuyeres, or the smelting zone. The oxygen in the blast combines 
with the fuel in the smelting column as soon as it enters the furnace, and 
it is here that the heat of combustion is liberated In iron blast 
furnaces the air is preheated before entering the tuyeres, but non- 
ferrous furnaces usually employ a cold blast. 

2 The diameter of a circular furnace, or the width of a rectangular 
furnace, is determined by the distance the air blast will penetrate the 
charge. It is essential that the rising gases be distributed evenly 
throughout the cross-section of the column. 

3. The smelting action in a blast furnace is essentially a reaction 
between the solids in the charge column and the stream of gas formed 
by the combustion at the tuyeres; the nature of this action will be 
determined by the make-up of the solid charge and the composition 
of the gases. In reduction smelting (iron, lead, and oxidized copper) 
an excess of coke is used and the combustion gases therefore contain 
large amounts of CO and the furnace is said to have a reducing atmos- 
phere; this plays an important part in reducing the oxides to the ele- 
mental metal. When a small amount of coke and an excess of air are 
used the combustion products contain free oxygen and the furnace has 
an oxidizing atmosphere; the gases then act to oxidize the solid matter 
in the charge. The relative amounts of fuel and air used must be 
adjusted according to the nature of the smelting done. 

4. Sufficient oxidizable material (carbonaceous fuel, sulfides, or both) 



138 SMELTING 

must be present before the tuyeres so that the heat generated will be 
sufficient to raise the temperature well above the formation temperature 
of the slag. Slag and either reduced metal or unoxidized sulfides 
(matte) become liquid in the smelting zone and trickle down to 
collect in the crucible. 

5. Part of the heat generated in the smelting zone is carried out as 
sensible heat in the molten products; the remaining heat is carried 
upward in the hot gases. A good deal of this heat is absorbed by the 
relatively cold solids in the upper part of the charge, so that the 
temperature of the solids rises as they descend in the furnace, and the 
temperature of the gases decreases as they rise toward the furnace 
top. In reducing smelting a part of this heat is absorbed by endo- 
thermic reactions such as (for reduction smelting of copper oxides) 

CuO + CO - Cu + CO 2 - 33,140 Cal 
Cu 2 + CO -> 2Cu + C0 2 ~ 28,040 Cal 

Also if carbonates are present either as ore minerals or as flux, they will 
decompose endothermically, viz.: 

CuC0 3 - CuO + CO 2 - 0,200 Cal 
CaC0 3 - CaO + C0 2 - 39,900 Cal 

The solids and gases at the top of the smelting zone should be cool; 
when the upper part of the column becomes abnormally hot, the con- 
dition is known as overfire. Overfire causes the charge to soften 
while still high up in the furnace and increases the formation of 
accretions which adhere to the furnace walls and interfere with the 
proper descent of the charge. 

6. The fact that blast furnace smelting is essentially the reaction be- 
tween a column of solids and an uprising current of hot gas imposes 
certain limitations on the physical properties of the charge. In the 
first place the charge must not contain much fine material, for two 
reasons: (1) A good part of the fine material would not stay in the 
column but would be immediately carried out of the furnace by the 
gas current, and (2) the fine material that did get into the column 
would tend to pack and form regions through which the gases would not 
pass, and thus the gas current would " channel " and leave portions 
of the charge cold and unsmelted. 

Also the solid particles on the charge must not be too large because 
the reaction with the gases can take place only at the surface, and 
large lumps would therefore smelt very slowly. 'The porous sinter 
produced by the Dwight-Lloyd machine makes ideal blast furnace 



THE COPPER BLAST FURNACE 



139 



feed because the pieces can be penetrated by the gases and also because 
the incipient fusion in the Dwight-Lloyd machine gives a " presmelted " 
or " predigested " material which smelts readily. Coke is the ideal 
blast furnace fuel because it is strong enough not to crush under the 
weight of the charge column and because its porosity aids in rapid 
combustion. Charcoal is occasionally used for blast furnace smelting; 
although porous, it is not as mechanically strong as coke and is 
generally much more expensive. Non-porous fuels such as wood or 
lump coal burn much too slowly to make suitable blast furnace fuel. 
Some experiments have been made in which powdered coal was blown 
in through the tuyeres, but this practice has not been adopted to any 
extent. 

With these general operations in mind, let us turn our attention to 
the copper blast furnace for smelting to matte this will be oxidizing 
smelting. Later we shall mention the reduction smelting of oxidized 
copper ores in the blast furnace. 




Launder to Converter 



To Slag Drain 

(Reproduced by permission from Ho/man and Hayward> Metallurgy of Copper, p 114, 

McGraio-Hill Book Co , New York, 1994) 

Fio. 21. Section Through a Copper Blast Furnace. 

The Copper Blast Furnace. Figure 21 is a cross-section of a copper 
blast furnace and Figure 22 shows another furnace in construction. 
Most furnaces have a height and width of the same order of magnitude 



140 SMELTING 

as those shown in Figure 22. These furnaces, however, may vary con- 
siderably in length. As a rule the width at the tuyeres will range from 
42 to 56 inches, and the length from 266 inches to 1044 inches. This last 
figure (87 feet) represents the length of a furnace used at Anaconda 
previous to 1918; this was the largest copper blast furnace in the world. 




FlQ. 22. Copper Blast Furnace Under Construction. 

These furnaces have vertical ends and sloping or boshed sides; the 
crucible is shallow and the slag-matte mixture is discharged through a 
raised spout which traps the blast and keeps it from blowing out of the 
taphole. The liquids usually flow continuously from the spout into 
the refractory-lined forehearth; slag overflows the forehearth and 
matte is tapped from the bottom (Fig. 21). 

The side walls of the furnace are made of hollow steel or cast-iron 
water jackets through which water is kept circulating. The pipes 
which circulate this cooling water can be seen in Figure 22. The 
tuyeres pass through the side walls and are connected by means of 
tuyere pipes to the large bustle pipe, which takes the air from the 
blowing engines. The top is closed and connected with the flue system 



CHEMISTRY OF MATTE SMELTING 141 

for the removal of the waste gases. Side doors in the upper part of the 
furnace are used for charging. The furnace shown in Figure 22 has 
the customary two tiers of water jackets. The water jackets are 
usually surmounted by a brick superstructure which contains the charg- 
ing doors and confines the gases until they enter the flue. 

A depth of charge column of 10 to 14 feet is commonly used, and the 
blast pressure will range from 10 to more than 60 ounces per square 
inch. The tuyeres usually have circular openings which range from 
2 to 6 inches in diameter, and there may be from 24 to 150 tuyeres to a 
furnace, depending upon the length of the furnace and the size of the 
tuyere openings. 

Chemistry of Matte Smelting. The reactions which take place in 
the blast furnace depend upon the nature of the charge, amount of 
coke used, and the volume of the blast. If it is desired to simply melt 
the charge down to matte and slag, after the manner of the reverbera- 
tory, then sufficient coke is used to provide the necessary heat, and 
the volume of the blast is regulated so that there is not quite enough 
oxygen present to burn all the carbon to C0 2 . This gives a slightly 
reducing atmosphere in the furnace, and except for the distillation of 
free-atom sulfur in pyrite and chalcopyrite, there is little or no 
sulfur loss, and the sulfides, gangue, and fluxes melt down in the 
smelting zone and form layers of matte and slag. 

If the blast supplies oxygen in excess of that required by the carbon 
in the fuel, the furnace atmosphere will be oxidizing and the oxygen will 
attack the iron sulfides present, thus: 

2FeS + 3O 2 -* 2FeO + 2S0 2 + 223,980 Cal 
2FeS 2 + 5O 2 -> 2FeO + 4S0 2 + 340,760 Cal 

The copper sulfides will not oxidize to any extent as long as some iron 
sulfide remains; if any copper sulfide were oxidized it would immedi- 
ately be converted back to sulfide by metathesis with the iron sulfide. 

The additional oxygen serves to oxidize the iron and cause it to 
enter the slag, and this serves to eliminate both iron and sulfur and 
hence raises the grade of the matte. For a given percentage of fuel on 
the charge, the grade of matte can be regulated within limits by 
regulating the volume of blast. 

In the smelting of massive pyrite ores, the practice of pyritic smelt- 
ing was developed to smelt copper ore in the blast furnace without the 
use of any carbonaceous fuel. This method was used at Mount Lyell, 
Tasmania, and other places where massive pyrite ores occurred. Py- 
ritic smelting has not been practiced for many years, but a brief dis- 



142 SMELTING 

cussion of the method will aid in an understanding of the more 
modern practice. 

Two things were essential to successful pyritic smelting (1) a 
" pyritic ore/' i.e., a massive pyrite ore containing copper in the form 
of disseminated chalcopyrite, and little gangue, and (2) a flux of almost 
pure silica. These materials were fed into the furnace and the iron 
sulfide burned to FeO, and the FeO then combined with the free Si0 2 
to form a ferrous silicate slag. Both the combustion and the slag- 
forming reactions were exothermic and together they supplied enough 
heat for smelting. 

In the pyritic furnace a column of infusible quartz extended from 
the bottom to the top of the smelting column; in the upper part the 
solid sulfides were present, but as the charge moved down in the 
furnace the sulfide melted and became oxidized and the FeO formed 
immediately reacted with the incandescent quartz to form slag. Thus 
the pieces of silica in the smelting zone were gradually eaten away by 
the corrosive action of the FeO, and as they disappeared they were 
replaced by more flux moving down in the smelting column. The slag 
and unoxidized sulfides (iron and copper) were withdrawn from the 
bottom of the furnace. 

The pyritic process was economical in that it required no ex- 
traneous fuel, but this apparent advantage was actually a weakness 
because it gave no flexibility to the process When coke is used as 
fuel it is possible to use more or less fuel as required, and, except for 
the minor effect of the coke ash on the slag, the composition of charge 
and flux need not be changed. On the other hand, in pyritic smelting 
the ore and flux served as its own fuel, and if the composition of thehC 
changed so that they could not provide enough heat for smelting, the 
process would not work. The bulk of the silica on the charge had to be 
present as " free " or uncombined SiOo because the heat of the reactions 
such as 

2FeO + Si0 2 -> Fe 2 SiO 4 + 22,000 C'al 

aided in the smelting. If the silica present in the gangue or flux were 
already combined with bases, then there would be no heat of com- 
bination, but these silicates would still have to be melted down. 

Consequently, as it became necessary to smelt ores which were not 
ideally suited for pyritic smelting, the practice was modified to the 
extent of adding some coke to the charge (0.5 to 6.0 per cent by 
weight) , and the method was then known as partial pyritic smelting. 
When the amount of coke used increases above 2 or 3 per cent of the 
weight of the charge, the process loses the characteristics of true 



TENNESSEE COPPER 143 

pyritic smelting and approaches the type of smelting that is used at 
present. In the modern blast furnace the principal fuel is coke; the 
oxidation of iron sulfides serves to raise the grade of the matte formed, 
and their combustion together with exothermic slag-forming reactions 
contribute considerable heat to the smelting operation, but these are 
not the principal sources of heat. It may be noted that one kilogram 
of FeS 2 , if oxidized to FeO and S0 2 and the resulting FeO combined 
with free silica, will yield about 1500 Cal of heat; a good grade of 
coke should have a calorific power of better than 7000 Cal per kilo- 
gram. Moreover, in pyritic smelting most of the " free-atom " sul- 
fur in pynte was distilled off in the upper part of the furnace and did 
not reach the smelting zone, where its heat of combustion would be 
useful. 

The slag and matte formed in blast furnace smelting do not differ 
significantly from the slags and mattes that we have discussed pre- 
viously. Following are brief descriptions of some blast furnace 
operations. 

Mount Lyell, Tasmania. 24 Smelting at Mount Lyell, Tasmania, 
began in 1896 and passed through various stages of development of 
pyritic and partial pyritic smelting to present simple smelting of raw 
and sintered concentrate, high-grade siliceous ore, and returned con- 
verter slag. 

The Mount Lyell blast furnace is 42 by 126 inches at the tuyeres 
and 13 feet 10 inches from tuyeres to feed floor. A blast pressure of 
20 to 30 ounces per square inch is used, and the furnace smelts about 
300 tons of new copper-bearing material per day. The charge is high 
in copper, and a 50 per cent matte is produced. Part of the flotation 
concentrate is mixed with flue dust and crushed limestone and is 
sintered on two Dwight-Lloyd machines; the remainder of the con- 
centrate is charged directly into the blast furnace. Analysis of the 
charge constituents and slag are given in Table 8. 

The converter slag must be allowed to solidify before it can be 
returned to the blast furnace; it cannot be added in the liquid form 
as is customary in reverberatory smelting. 

Tennessee Copper. 25 The present-day treatment of the massive 
sulfide ores at Copperhill, Tennessee, is interesting in that the ores 
are treated to recover the contained iron, sulfur, copper, and zinc in 
the form of marketable products. This practice is a great improvement 

24 Metallurgical Operations at Mount Lyell, Tasmania, Eng. and Min Jour , Vol. 
130, No. 1, p. 12, 1930 

25 Tennessee Copper Works Toward Maximum Economy: Eng. and Min. Jour., 
Vol. 138, No. 10, p. 40, 1937 



144 



SMELTING 



TABLE 8 
MOUNT LYELL 



PmHiipf 








Per 


cent 










Cu 


S 


Fe 


FeO 


A1 2 3 


Si0 2 


CaO 


BaSO 4 


Sinter 


20.5 


12.6 


27.6 






17.8 






Raw concentrate 


18.3 


33.8 


28.2 


.... 


4.3 


10.2 


.... 




High-grade North Mount 
Lyell ore 


17.0 




9.8 


.... 


4.1 


47.2 


.... 


6 3 


Slag 


45 






53 1 


5.89 


33.4 


3 13 


a O 70 



BaO. 



The relative amounts of each material on the charge are as follows: 



Sinter 

Raw concentrate 

Converter slag 

North Lyell ore 

Coke, about 9 per cent of solid charge 



Pounds 
1050 
1850 

800 

300 

325-350 



on the previous method, which employed principally blast furnace 
smelting; in 1916, there were seven blast furnaces used and today only 
one remains. The principal function of this furnace is to supply 
matte for the converter, and all the copper concentrates are smelted 
in the converter. We shall take up the direct smelting of concentrates 
in the converter in the next chapter. If it were not for the fact that 
some liquid matte is needed for the converter operation, blast furnace 
smelting would probably be abandoned completely; and a good deal 
of the iron which now passes into the blast furnace slag could be 
thrown into the iron concentrates. This blast furnace may eventually 
be replaced by a reverberatory furnace. 

The crude ore is first subjected to hand picking, and the high-grade 
material removed serves as blast furnace feed. The mill feed is then 
treated by selective flotation to produce concentrates of the following 
average composition: 

Iron concentrates 54.0 per cent Fe, 40.0 per cent S 
Zinc concentrates 50.0 per cent Zn 

Copper concentrates 20.0 per cent Cu, 35.0 per cent Fe, 35.0 per 
cent S, 3.5 per cent Si0 2 , 2.0 per cent Zn 

The zinc concentrates are sold to a zinc smelter and the copper con- 
centrates are smelted in the converter which treats the blast furnace 
matte The iron concentrates are first roasted down to 6.5 per cent 



CONISTON 145 

sulfur in Wedge roasters; the calcine is then mixed with coke breeze 
and sintered on Dwight-Lloyd machines to give an iron oxide sinter 
containing only 0.05 per cent sulfur. This sinter is then sold for smelt- 
ing in the iron blast furnace. 

All S0 2 in the gases goes to a sulfuric acid plant, and of the gas 
delivered, 60 per cent comes from the roasters, 25 per cent from the 
blast furnace, and 15 per cent from the converter. Actually it is the 
requirements of the acid plant which determine the smelter practice. 

The blast furnace used at Copperhill is 22 feet by 4 feet 8 inches at 
the tuyeres and runs with a 16-foot column and 30-ounce blast. The 
matte produced contains approximately 12 per cent copper and 24 
per cent sulfur; the furnace has a maximum capacity of 1000 tons of 
charge per 24 hours but normally treats about 650 tons made up ap- 
proximately as follows: 

Tons 

Ore from Burra Burra and Eureka mines 500 
Ore from Fontana mine 50 

Quartz flux 60 

Coke 40 

The coke used makes up about 6.0 per cent of the charge. The 
average analysis of the charge is Cu, 2.5 per cent; Fe, 33.0 per cent; 
S, 25.0 per cent; Si0 2 , 18.0 per cent. The Fontana ore is richer in 
copper (about 7.0 per cent) than the other ores, but it contains talc, 
which makes it harder to smelt. 

Converter slag is not resmelted, but the molten converter slag is 
poured into the forehearth of the blast furnace where the entrained 
matte settles out. 

Coniston. 26 The Coniston smelter of the International Nickel Com- 
pany (near Sudbury, Ontario) employs four blast furnaces to treat 
coarse magnetic ore from the Frood mine and both coarse and fine 
magnetic ores from the Creighton mine. These ores are massive 
sulfides high in copper and nickel and low in silica. The furnaces 
make a copper-nickel matte whose grade varies from time to time 
depending upon the ore being smelted. 

Coarse ore is charged directly to the blast furnace, and fine ore, flue 
dust, and limestone are sintered on six Dwight-Lloyd machines, the sin- 
ter then going to the blast furnaces. The blast furnaces are 50 by 240 
inches at the tuyeres and have settlers or forehearths 18 feet in diameter. 
The furnace water jackets extend down only to the top of the crucible; 

26 Canadian Min. Jour., p. 683, November 1937. 



146 SMELTING 

the furnace crucibles (" bottoms " of the furnaces) are lined with mag- 
nesite brick, and the settlers also. The breast jackets and spouts on 
the furnaces are made of cast iron and are water cooled. Molten con- 
verter slag is returned to the blast furnace settler. 

These three examples illustrate the use of the blast furnace for 
matte smelting. Blast furnaces are seldom thus employed and then 
only under special conditions; the bulk of the copper-bearing feed 
(finely divided concentrate) which reaches most copper smelters is not 
suited for blast furnace smelting. The blast furnace is cheaper to con- 
struct than a reverberatory and for a small installation may prove 
cheaper to operate than the reverberatory when everything is taken 
into account. For large installations, however, the reverberatory is 
the standard. 

ELECTRIC SMELTING 

The smelting of copper concentrates for matte in electrically heated 
furnaces has been practiced in Europe, but in the Western Hemisphere 
the cost of electric power has been too high to compete with fuel 
heating. Electric smelting has been confined to a few remote regions 
of the world where water power is abundant and fuel is expensive. . 
Technically, the electric smelting furnace offers many interesting 
features not found in the fuel-fired reverberatory, and it may well be 
that in the future more use will be made of this technique. 

The discussion which follows is taken from an article by M. Sem, 27 
of Oslo, Norway, and deals with the use of electric smelting in Norway 
and Finland. 

Figure 23 is a diagrammatic sketch showing one of the early Westly 
smelting furnaces, and this illustrates the principles involved. Actually 
the furnace is simply a melting furnace, and heat is supplied by the 
current carried in on the carbon electrodes which pass through the 
roof and dip into the slag bath. The heat is generated by the resistance 
the bath offers to the heavy amperage current brought in by the 
electrodes; as the smelting progresses, the electrodes slowly burn 
away, and provision is made to lower them as required. The charge 
enters through the roof of the furnace and melts down to slag and 
matte. The slag can be tapped from either end, and the matte collects 
in a sump at one end of the furnace, from which it can be tapped as 
required. 

The chemistry of the electric smelting process is essentially the same 
as in reverberatory smelting; matte and slag are formed in the same 

27 Sem, M., Electric Smelting with the Westly Furnace: Eng. and Min. Jour., 
Vol 140, No. l,p.47, 1939. 



ELECTRIC SMELTING 



147 



way, and the interaction of the various components of the charge results 
in the elimination of more or less of the sulfur as S0 2 gas. The 
principal drawback to the use of electric smelting is the cost of electric 
power, as we have noted, and the method can only be used at present 
where the cost of power is unusually low. Another objection to electric 
smelting has been the fact that most of the furnaces were small and 



Charge opening 
Stag may be tapped 
at either end 




(Sem, Eng and Min. Jour., Vol. 140, No. l t p. 47, 1959) 
FIG. 23. Diagrammatic Sketch of the Early Westly Electric Smelting Furnace. 

had rather low capacities; this second disadvantage has been largely 
overcome, and it is now possible to build electric furnaces of such a 
size that they have a capacity comparable with that of the standard 
reverberatory. 

The largest closed electric smelting furnace in the world was put 
into operation at Imatra, Finland, in 1936. This furnace has an 
outside diameter of 10 meters and is heated by a 3-phase current 
carried on three Soderberg electrodes each 1.4 meters in diameter and 
weighing about 15 tons. Power consumption ranges between 500 and 
600 kwhr, and electrode consumption is from 2 to 3 kg of electrode per 
ton of charge. The furnace is lined with magnesite and has separate 
tapholes for slag and matte; the slag may flow continuously, and the 
matte is tapped into a ladle through water-cooled tapholes. 

The furnace is charged mechanically from overhead hoppers and 
handles from 240 to 250 tons of cold charge per day plus the molten 
converter slag. The raw material is a concentrate containing about 
20 per cent copper; part of the concentrate is roasted and part is 



148 SMELTING 

charged raw. The matte produced has an average grade of about 
50 per cent copper; the slag will assay from 0.3 to 0.6 per cent copper, 
depending upon the grade of the matte. 
Some of the advantages of electric smelting may be listed as follows: 

1. The fact that fuel is not used means that the thermal efficiency 
of the furnace is much greater than that of the reverberatory. There 
is no great volume of high-temperature combustion gases passing out 
of the furnace to carry away as much as 50 per cent of the heat gen- 
erated. Of course most of this heat is abstracted by waste-heat boilers 
in modern reverberatories, but the heat is not confined to the furnace 
itself as it is in the electric furnace. 

2. The absence of the large volume of combustion gases which 
sweep through a fuel-fired reverberatory gives the electric furnace a 
quiet atmosphere which greatly reduces dust losses and the corrosion 
and abrasion of refractories caused by the impact of a current of 
hot dust-laden gas. 

3. There are no combustion gases to dilute the S0 2 gas evolved from 
the smelting charge, and hence there is a much smaller volume of gas 
to handle and it is much richer in S0 2 . This is a very important 
factor when the gas is to be treated to remove the sulfur. 

4. Present-day furnaces show reasonably low power and electrode 
consumption, the slag losses are low, and the operation is simple 
and reliable. 

All the smelting methods we have considered up to this point have 
been processes of matte smelting. We shall now proceed to consider 
two smelting methods which yield metallic copper directly smelting 
of native copper concentrates and the reduction smelting of oxide 
copper ores and concentrates. 

SMELTING OF NATIVE COPPER 

The Lake Superior district is the only place in the world where 
native copper has been found in great abundance, and it is here that 
we find the only application of native copper smelting. The following 
discussion is taken from an article describing the smelter of the Calumet 
and Hecla Consolidated Copper Company on Torch Lake, Michigan. 28 

The concentrates coming to the smelter consist of metallic copper 
plus some gangue minerals. There is also a certain amount of copper 
oxide (from the ammonia leaching plant) which is smelted with the 
native copper. Except for the reduction of the copper oxide precipitate 

^Lovell, E. R., and Kenny, H. C., Present Smelting Practice: Mining Cong. 
Jour., p. 67, October 1931. 



SMELTING OF NATIVE COPPER 149 

the operation is simply one of melting the copper and fluxing the 
gangue minerals. 

Large pieces of mass copper from the mines and mills go directly 
into the furnace. Gravity and flotation concentrates are stored in 
bins until they are needed; then they are bedded and mixed in the 
proper proportions. The gravity concentrates from the conglomerate 
lode have a ferruginous gangue; the amygdaloid concentrates have a 
gangue containing silica and some alumina; and all the flotation con- 
centrates have a gangue which is siliceous and high in alumina as 
well. The various concentrates are intimately mixed in such propor- 
tions that the gangue minerals are self-fluxing and about 5 per cent 
of coal or coke screenings is added to the mixture. 

Two such mixtures are made up and stored in separate bins. One 
is high grade and contains about 75 per cent copper, the other is 
low grade and averages about 40 per cent copper. The amount of 
each mixture to be added to the furnace charge depends upon the im- 
mediate demand for copper from the melting furnace. Limestone flux 
is added with the mass copper. The copper oxide is stored in a separate 
bin and is mixed with 8 per cent of coal screenings before being charged. 

The melting furnace is a reverberatory furnace resembling the 
furnaces used in matte smelting. All furnaces are equipped with 
water-cooled cast-iron side plates to prevent break-outs. Skewbacks 
and charge-hole jackets are made of deoxidized copper and are also 
water cooled. The slag is skimmed through tapping openings on one 
side of the furnace, and a tapping slot for copper is on the opposite 
side; there is one slag tap near each end of the furnace, and the 
copper taphole is located near the burner end. Furnace foundations 
are of solid concrete, and the furnace bottom is a silica-brick inverted 
arch 18 or 20 inches thick; under the bottom proper is another inverted 
silica-brick arch. The use of burned-in silica or sand bottoms has not 
been successful. The furnace is fired with pulverized coal and is 
charged through holes in the center of the roof. 

It is customary to charge a certain number of holes every 2 hours, or 
at longer intervals, depending upon the rapidity of melting at various 
points along the furnace. While the charging is going on, slag is being 
drawn off more or less continuously, starting as soon as the slag be- 
comes sufficiently fluid usually 4 to 6 hours after the first charge. 
The furnace atmosphere is kept slightly reducing, and this, together 
with the coal mixed with the concentrates before charging, serves to 
prevent oxidation of the copper and consequent slag loss. Approxi- 
mately 14 hours before the copper is to be tapped from the furnace, 
charging is stopped and the piles of concentrate are allowed to melt 



150 SMELTING 

down. Before the piles are flat, air is blown below the surface of the 
bath for 4 to 6 hours to assist in bringing up unmelted material from 
the bottom; this tends to oxidize the bath, but as long as a blanket of 
fine coal remains on the bath there is little danger of excessive slag 
losses due to oxidation of the copper. No slag is tapped while the air 
is being blown, but after the blowing is stopped the slag is allowed time 
to separate, and tapping is then continued. From 2 to 4 inches of 
slag is allowed to remain on the surface of the bath, and the copper 
is tapped from beneath it. 

Molten copper is not as easily handled as molten matte its melt- 
ing point is higher, and its specific heat is so low that unless the 
molten copper is heated well above its melting point it will solidify 
and form skulls in the ladles and launders. For this reason it is the 
custom to melt practically the entire charge before tapping copper, 
because it is almost impossible to raise the molten copper above its 
melting point if any unmelted copper remains in the furnace. The 
slag is granulated and pumped to the waste slag dump, and the copper 
flows from the melting furnace directly to the fire refining furnace. 
The oxide copper precipitate mentioned above is usually charged in 
this refining furnace rather than in the melting furnace. 

The plant contains two melting furnaces and two refining furnaces. 
Usually only one of each is in service and the others are kept ready for 
use whenever the necessity arises. The average daily capacity of the 
melting furnace is about 280 tons of concentrates, which will produce 
about 120 tons of slag; the holding capacity of the furnace is about 
500 tons of molten copper. At the maximum firing rate the furnace 
will consume about 75 tons of coal a day. Analyses of slag and 
melting furnace copper are: 



Melting Furnace Copper 

Per Cent 
Cu 98.7 

Fe 1.0 

S 0.2 

As 0.04 



Slag 

Per Cent 
SiO 2 42.5 

FeO 30.0 

A1 2 O 3 13.0 

MgO 2.3 

CaO 8.0 

Cu 0.60 



The slag would not be considered desirable in a matting furnace 
because of the high A1 2 3 content, although the silicate degree is 
about right (1.4). It is sufficiently fusible, however, at the higher 
temperatures employed in the copper melting furnace. 

It was previously the custom to resmelt the slag in a blast furnace 



SMELTING OF OXIDIZED COPPER ORES 151 

to remove the copper, but, as firing with pulverized coal and mixing 
coal with the charge were practiced, the copper content of the slag is 
held low enough so that resmelting is not required. 

SMELTING OF OXIDIZED COPPER ORES 

Introduction. The smelting of ores containing the oxidized copper 
minerals rather than sulfides differs from the smelting of sulfide ores 
in two respects (1) the process is one of reduction smelting and 
(2) the product is metallic copper and not matte. The copper pro- 
duced by reduction smelting is quite impure and is often called black 
copper. The ore minerals may be oxides (cuprite, tenorite), carbon- 
ates (malachite, azuritc) , or silicates (chrysocolla) ; the carbonates 
break down to copper oxide and C0 2 in the furnace, and the copper 
silicate is decomposed by more active bases (CaO, FeO) to yield 
copper oxide and calcium or iron silicate. Thus all these compounds 
eventually behave like copper oxide, and the essential chemistry of the 
process involves the reduction of the oxide to metallic copper. 

As we shall see directly, the smelting of oxidized copper ores has 
never been as satisfactory as matte smelting. The most successful way 
of treating oxidized copper ore is by leaching, and today most of 
the ores from the great oxide ore bodies are being treated by leaching. 
One exception is the high-grade oxide ore from Katanga in the Belgian 
Congo, and here also a large leaching plant has been built to treat 
certain classes of ore. It is likely that leaching will some day displace 
smelting completely for all classes of oxide ore in the Katanga district. 

Early Arizona Practice. In the early days in Arizona, high-grade 
oxidized ores were smelted directly in blast furnaces. These furnaces 
were water-jacketed throughout; some were rectangular in cross- 
section, but most of them were circular furnaces. Some analyses of 
smelting ores, slags, and the black copper produced are given in Table 9. 

This type of smelting had two principal disadvantages; the copper 
loss in the slag was very high and the copper produced was quite im- 
pure and required considerable refining. Eventually this practice 
was abandoned, and the high-grade oxide ores were mixed with sulfide 
ores and concentrates and smelted to matte a procedure which was 
also applied to some of the high-grade slags which had been sent to the 
dump from the blast furnaces. Low-grade oxide ore was treated 
by leaching. 

Note that the slag losses were high always well over 1.0 per cent 
and in one smelter almost 4.0 per cent CuO. Part of the copper was 
present as grains of metallic copper, but most of the copper in the slag 
was there as copper silicate caused by the slagging of copper oxide be- 



152 



SMELTING 



TABLE 9" 
ANALYSES OF ARIZONA OXIDE COPPER ORES 



Locality 


Cu 


Si0 2 


FeO 


MnO 


CaO 


Longfellow 


38.80 


11 15 


10 40 






Do 


21 67 


17 25 




7 43 




Do. 


17.17 


26 80 


13 76 


7.49 


.... 


Coronado 


21.95 


48 90 


12 09 






Do. 


11 17 


67 00 


8 88 






Old Dominion 


15 17 


35 3 


28 7 




22 2 



ANALYSES OF ARIZONA BLAST FURNACE SLAGS 



Smelter 


Cu 


CuO 


SiO 2 


FeO 


MnO 


CaO 


MgO 


A1 2 3 


S 


Copper Queen 


2 10 




24 67 


44 85 


39 


10 92 


1 75 


15 57 


28 


Do. 


15 


1.02 


30 06 


53 36 


11 10 










Detroit 






34 34 


32 27 


8 05 


10 13 


2 30 


11 64 




Do. 


1 82 




29.50 


37 08 


1 13 


9 02 


7 44 


14.07 


30 


Prince 


1 64 




27 16 


34 62 


49 


17 42 


3 51 


14 70 


33 


Old Dominion 




3 76 


27 23 


51 30 


1 65 


5 14 


2 54 


5 22 




United Verde 


18 


2 59 


35 79 


37 89 




12 98 


75 


8 29 




Bisbee 


1 32 




28 


29 




9 




27 





ANALYSES OF ARIZONA BLACK COPPER 



Smelter 


Cu 


S 


Fe 


As 


Sb 


Bi 


Insoluble 


Prince 


95 00 


44 


4.23 








51 


Old Dominion 


98 91 


64 


12 


057 


008 


010 


0.065 


Do. 


98 27 


60 


73 


039 


019 


trace 


060 


Do. 


98 24 


53 


80 


051 


021 


006 


007 


Do 


97 52 


69 


97 


052 


014 


trace 


180 



Hofman, H O , and Hayward, C R , Metallurgy of Copper, p 247, McGraw-Hill Book Co , New 
York, 1924 

fore it could be reduced. Some of the analyses indicate the amounts of 
the copper present as elemental copper and as oxide; where only the 
copper is given this refers to the total assay and most of this copper 
is present as the oxide. 

Practice at Katanga. 29 The ores of Katanga are oxidized ores with 
a gangue that is highly siliceous and low in iron ; most of the gangues 
are quartzitic or schistose, a few ores have a dolomite gangue. The 
principal ore minerals are malachite and chrysocolla with minor 
amounts of other oxidized copper minerals. The ores vary widely 

29 Roger, E., La Metallurgie du Cuivre et du Cobalt au Katanga : Cong. Internat. 
des Mines, de la Metallurgie, et de la Geologic Apphquee (Sect. M6t.), pp. 441-456, 
6me sees., Liege, June 1930. 



PRACTICE AT KATANGA 153 

in grade. Much ore is available which contains more than 15 per cent 
copper, and other ores contain as little as 4 per cent copper. 

The Union Miniere du Haut Katanga operates the mines and metal- 
lurgical plants of the district, and we shall discuss the metallurgical 
operations in the order of their adoption. 

In considering the methods to be used in treating these ores it was 
decided to use reduction smelting in water-jacketed blast furnaces; 
these were used on the rich ore from the Mine de TEtoile, which ran 
about 15 per cent copper. 

In 1911 there was one such blast furnace in operation in the smelter 
at Luburnbashi, from 1911 to 1914 two more were constructed, and in 
1915 another extension was made which brought the total to seven. 
The first three furnaces constructed consisted of a single stage of 
water jackets surmounted by a brick superstructure; the dimensions 
at the crucible and the tuyeres were 1 20 by 4 88 meters (47 2 by 192 
inches). There were 8 water jackets (4 2 by 6 meters) on each side, 
and each was pierced by two tuyeres, making a total of 32 tuyeres per 
furnace. The tuyere openings were each 127 square centimeters in 
cross section (equivalent to a circular opening about 5 inches in 
diameter). The last four furnaces had two stages of water jackets 
and a crucible 1 12 by 610 meters (44 1 by 240 inches) ; the total 
height of the water jackets was 4 35 meters, and the charging floor was 
082 meter above the top of the upper btage of water jackets These 
furnaces each had 40 tuyeres 127 square centimeters m cross-section 
two in each of the 20 lateral water jackets on the lower stage. These 
furnaces had internal crucibles and employed a blast pressure of 80 
to 110 millimeters of water. 

Later, changes were made in the size and shape of the furnaces; 
the section at the tuyeres was increased to 1 57 by 6 10 meters (620 
by 240 inches), and the number of tuyeres was cut in half. The new 
tuyeres had an opening of 190 square centimeters (about 63 inches in 
diameter), but this increase was not enough to give the same total 
tuyere area as before. This reduction in tuyere area had a beneficial 
effect on the smelting rate. 

The reactions which take place in these furnaces are relatively 
simple; in the upper part the charge is dried, and carbonates in the 
ore and flux are decomposed into CO 2 and the metallic oxides; reduction 
of oxides takes place farther down in the furnace, and in the smelting 
zone above the tuyeres the molten slag forms; rectal and slag then 
trickle down to collect in the crucible, from which they are tapped. 
The copper oxides are reduced to the metal together with some of the 
oxides of iron and cobalt (if present). The resulting metal contains 



154 SMELTING 

about 97 or 98 per cent copper. A small amount of matte forms 
because there is a little sulfur on the charge and this comes principally 
from the coke used as fuel. 

In smelting these ores it was necessary to use a temperature high 
enough to insure reduction of the copper oxides, but if too much fuel 
were used the temperature would be so high and the reducing action 
so strong that too much iron would be reduced from the iron oxides 
on the charge. It was therefore necessary to adopt a type of slag 
which would have a formation temperature between 1150 and 1300 C. 
The slags which have been used in these furnaces are sesquisilicates of 
lime and iron containing from 40 to 45 per cent SiO 2 . Slags which 
have been successfully used over a period of years have compositions 
between the following limits: 





Per Cent 


SiO 2 


40 to 45 


A1 2 O 3 


G to 8 


FeO 


14 to 19 


CaO 


14 to 21 


MgO 


5 to 8 



Since the gangue minerals are generally siliceous, basic fluxes are 
required; limestone and dolomite are used for the lime and ohgonite 
and hematite to supply the iron. 

The annual capacity of these blast furnaces is about 1,000,000 tons 
of charge; the annual production of metal from the blast furnaces 
from 1918 to 1927 is listed below. 

TABLE 10 

ANNUAL PRODUCTION, IN METRIC TONS, FROM KATANGA 
BLAST FURNACES 



1918 


20,239 


1923 


56,220 


1919 


23,019 


1924 


79,697 


1920 


18,962 


1925 


81,081 


1921 


30,464 


1926 


72,328 


1922 


43,362 


1927 


72,621 



Note that the total output in 1925 was more than four times the 
1918 production. This was made possible by changes in the furnace 
design and by the selection of higher grade material for the furnace 
charge. Some of the important facts about the smelting during two 
different periods are given in Table 11. 

The question of copper losses in the slag has been a matter of much 
concern, as the blast furnace slags carry from 1.5 to 2.5 per cent 



PRACTICE AT KATANGA 



155 



copper. Efforts to decrease the slag losses have followed two lines 
(1) to reduce the slag volume and (2) to reduce the copper assay 
of the slag. 

The first of these has met with considerable success, and the method 
employed has been simply to select higher grade feed for the furnaces. 
This not only gives a greater yield of copper per ton of ore smelted 
and requires less coke but, as there is less gangue on the charge, less 
flux is required and the slag volume is diminished accordingly. 
Table 11 shows that the weight of flux required has been cut in half; 
the weight of the slag produced was reduced from 120 per cent of the 
weight of the ore to 80 to 95 per cent. 

TABLE 11 
BLAST FURNACE SMELTING AT KATANGA 





1911-1920 


Since 1928 


Per cent of Cu on charge 


8 to 9% 


14 to 15% 


Tons of total charge (ore and flux) per hour 


10 7 tons 


20 to 24 tons 


Tons of ore per ton of produced metal 


8 4 tons 


5 tons 


Weight of flux ; per cent of weight of ore 


70 to 90% 


40 to 50% 


Weight of coke; per cent of total charge 


24% 


17 to 20% 


Tons of coke per ton of metal produced 


3 5 tons 


1 5 tons 


Tons of total charge per ton of metal produced 


14 5 tons 


7 6 tons 



Efforts to lower the copper content of the slag have not been so 
successful. At one time forehearths or settlers were used on the 
theory that this would diminish slag losses by permitting suspended 
globules of copper to settle out. The forehearths did not give the 
desired effect, however, because most of the copper in the slags was 
chemically combined as copper silicate and was not present as 
mechanically entrained copper. After several trials, the use of fore- 
hearths was abandoned. Better reduction of copper could be attained 
by using more coke and thus having a more vigorous reducing action; 
but as it was necessary to have considerable iron oxide on the charge to 
insure a fusible slag, this procedure would reduce too much metallic 
iron. The use of powdered coal blown through the tuyeres was also 
attempted in order to secure better reduction of the copper; but these 
experiments were not successful, and the practice was abandoned. 

In general these blast furnaces are operated to give a high tonnage 
of metal containing 96 to 98 per cent copper and as free as possible from 
iron and cobalt. The total slag loss does not exceed 10 per cent of 
the copper on the charge, and it is not economical to try to cut this 
loss at the expense of furnace capacity or grade of copper produced. 

In spite of the large output of copper from the blast furnaces it was 



156 SMELTING 

realized that blast furnace smelting was not suited for all types of 
ore in the district. Economical blast furnace smelting demanded 
coarse high-grade ores to produce copper with the minimum slag 
losses and coke consumption and maximum furnace capacity. Accord- 
ingly, other means had to be developed to take care of (1) ores too 
low in grade for the blast furnace and (2) rich ores that were too 
finely divided to go directly to the blast furnaces. 

Sintering. The fine ores and concentrates which were continually 
increasing in amount led to the installation of two Dwight-Lloyd sin- 
tering machines between 1923 and 1925. The charge to these machines 
was mixed with 9 to 11 per cent of its weight of coke breeze and 
sintered to nic^ke blast furnace feed. These two machines require just 
about as much coke breeze as is obtained by screening the coke for the 
blast furnaces, and no more sintering machines have been installed 
Fine ores and concentrates which are not wintered are smelted in re- 
verberatory furnaces. 

Concentration In Chapter I we have given a brief description of 
the Panda concentrator and have noted that both gravity and flotation 
concentrates are produced. These concentrates are then either smelted 
in reverberatory furnaces or sintered and charged to the blast furnaces. 
In addition to the more or less standard methods of concentration the 
Union Miniere has developed a process of reducing at low temperature 
copper minerals in crushed ore to metallic copper and then removing 
the metallic copper by flotation. On certain low-grade ores this 
method gives a concentrate assaying 65 per cent copper and 0.5 per 
cent tails, with a recovery of 90 per cent. 30 

The production of the two mills of the Union Miniere in 1936 was 
as follows: 31 

Panda mill : 

11,450 tons of 33 2 per cent gravity concentrates. 

50500 tons of 350 per cent gravity concentrates 

71,800 tons of 298 per cent flotation concentrates. 

Prince Leopold works : 

63,700 tons of 30 5 per cent concentrates. 

Reverberatory Smelting. The constantly increasing supply of finely 
divided concentrates at Katanga led to the construction in 1922 of an 
experimental reverberatory furnace at the Lubumbashi smelter. Two 
years of work with this furnace demonstrated the feasibility of smelting 

30 Roger, E , op cit , p 449. 

31 Minerals Yearbook, 1938, p. 103, U. S Bur. Mines. 



PRACTICE AT KATANGA 157 

oxide ores and concentrates in the reverberatory. In 1927 a new 
smelter containing four reverberatory smelting furnaces was put in 
operation; the new smelter was located near the Panda concentrator. 

These furnaces are 33 by 6 meters (108 by 19.7 feet), interior di- 
mensions, and are of silica brick construction throughout. Firing is 
done with pulverized coal and there are five burners to each furnace ; 
combustion air is preheated to 180 C by means of the heat left in the 
furnace gases after they pass through the waste-heat boilers. 

Fluxes are crushed to 6 millimeter size and thoroughly mixed with 
the copper concentrates and ores together with about 7 to 10 per cent 
by weight of fine coal; this serves as the reducing agent for the 
copper oxide. The charge is dried more or less completely in 8 Wedge- 
type 5-hearth furnaces fired with powdered coal burners. When flota- 
tion concentrate is being smelted the charge must not be completely 
dried or the dust losses in the furnace are excessive, and the charge 
piles in the furnace tend to slide down into the bath. 

Side-charging is employed, with the charge entering through holes 
in the roof and piling up along the sides of the furnace. The reduced 
copper and slag melt down and form a pool in the furnace from which 
they can be drawn as desired. It is necessary that the material shall 
not slide off the piles and float on the bath if this happens some 
of the unsmelted charge may be carried off with the slag. 

The slags formed are slightly more siliceous than those produced in 
the blast furnaces. Better settling conditions in the furnace permit 
the use of a more siliceous (and more viscous) slag, which cuts down 
on the amount of basic flux needed. The reverberatory slags are also 
somewhat lower in copper than the blast furnace slags when copper 
of similar quality is being produced. Furnace slag is granulated in 
water. 

Table 12 gives some of the data characteristic of reverberatory prac- 
tice at Katanga. 

Some of the advantages of reverberatory smelting over blast furnace 
smelting at Katanga are as follows: 

1. It permits the direct treatment of fine ores and concentrates which 
could not be smelted in the blast furnace without preliminary sintering. 

2. Slag losses are less, and the recovery of metal is usually higher in 
the reverberatory than in the blast furnace. 

3. Although the total consumption of fuel is greater than in blast 
furnace smelting, the coal is sufficiently cheaper than coke so that the 
fuel cost per ton of charge smelted is lower for reverberatory smelting 
than for the blast furnace. Moreover, about 35 to 40 per cent of the 
heat of the combustion coal is recovered by the waste-heat boilers. 



158 



SMELTING 



TABLE 12 
REVERBERATORY PRACTICE AT KATANGA 





Experimen- 
tal Furnace 


Panda Smelter 




1925 


1928 


Sept 1929 


Smelting rate, tons per hour 


7.9 tons 


8.4 tons 


11.2 tons 


Copper content of ore and concentrate, per 








cent 


24 80% 


27 79% 


29% 


Copper content of charge, per cent 


16% 


19% 


23% 


Coal, per cent of charge 


45 80% 


40% 


31 7% 


Coal, tons per ton of metal produced 


3 20 tons 


2 40 tons 


1 45 tons 


Daily production of metal; tons per furnace 


28 tons 


35 tons 


60 tons 



4. Blast furnace smelting demands selected ores of 15 to 20 per cent 
copper, and such a process requires (for coke manufacture and other 
purposes) about 2.5 tons of coal per ton of copper produced from 20 
per cent ore. The concentrator and reverberatory furnaces, on the 
other hand, will operate on lower grade ores and require only about 1.7 
tons of coal per ton of copper produced; also part of the heat is con- 
verted into steam power which can be used in mill and smelter. 

Leaching. In addition to the pyrometallurgical operations, the 
Union Mmiere has installed a large copper leaching plant. We shall 
have occasion to discuss this in a later chapter. 

OTHER SMELTING METHODS 

We shall mention here two papers dealing with methods of copper 
smelting; these have not yet had commercial application, but they 
indicate other approaches to the ideal which is back of such processes 
as pyritic smelting and direct smelting in converters, i.e., to use as 
advantageously as possible the heat of combustion of the sulfide min- 
erals, or to make the sulfides smelt themselves (autogenous smelting). 

Shaft Roasting and Reverberatory Smelting. 32 In the spring of 
1931 an experiment was made at Anaconda in which a down-draft 
roasting shaft (essentially a flash or suspension roaster) was mounted 
over a small reverberatory so that the hot calcine discharged directly 
into the smelting furnace. The hearth of the reverberatory was 
3 by 21 feet and the roasting shaft was 3 feet square and had an effective 
height of 20 feet. The shaft discharged through an opening in the 

32 Laist, Frederick, and Cooper, J. P., An Experimental Combination of Shaft 
Roasting and Reverberatory Smelting: Am. Inst. Mm. Eng. Trans., Vol. 106, 
p. 104, 1933. 



AUTOGENOUS SMELTING 159 

center of the roof about 7 feet from the burner wall. Dry concentrate 
mixed with the necessary flux constituted the feed to the roaster. 

When operated without the shaft, on calcine at a temperature of 
492 F, the reverberatory smelted 16,678 pounds per 24 hours with an 
oil consumption of 797 gallons. The roaster-reverberatory combination 
smelted 40,552 pounds of cold dry concentrate per hour with an oil 
consumption of 262 gallons. After making due allowance for the 
calcination ratio (change in weight due to roasting) of the concen- 
trates, it may be said that the use of the shaft increased the capacity 
of the smelting furnace 2 l /> times and decreased the fuel consumption 
by 60 per cent These are results obtained fnrn 30 8-hour shifts 
with the reverberatory alone and 30 8-hour shifts with the roaster- 
reverberatory combination. 

Laist and Cooper are of the opinion that such an installation on a 
standard reverberatory, with the shaft roasters dropping calcine 
through the roof along the side walls of the reverberatory, might 
reasonably be expected to increase the smelting capacity from 1% to 
2 times without any material increase in fuel consumption. Other 
factors might weigh against the practical use of such a combination ; 
for example, in certain cases the copper content of the concentrates 
will be so high that all the sulfur is needed for the matte, and such 
concentrates could not be roasted before smelting. 

Autogenous Smelting. In an article published in 1936 Mr. T. E. 
Norman 33 has discussed the theoretical aspects of smelting copper 
concentrates without the use of extraneous fuel. It appears that the 
maximum temperature attained in ordinary flash roasting in air 
is not high enough for the direct formation of liquid matte and slag 
from the calcine; this is because the large volume of inert nitrogen in 
the combustion air absorbs too much of the heat. If, however, the 
suspension roasting is carried out in an atmosphere of 40 to 95 per 
cent 2 (air contains 21 per cent 2 by volume), then there is less 
material to absorb the heat evolved and a higher " flame temperature " 
is attained high enough that properly fluxed concentrates would 
issue from the roaster not as solid particles of calcine but as drops of 
molten sulfides and slag together with superheated particles of acid 
and basic oxides which would react to form liquid slag at the first 
opportunity. 

The bulk of this article consists of an extended consideration of 
the production and distribution of heat in copper roasting and smelting, 

33 Norman, T. E., Autogenous Smelting of Copper Concentrates with Oxygen- 
Enriched Air: Eng. and Mm. Jour , Vol. 137, No. 10, p. 499, 1936, and Vol. 137, 
No. 11, p. 662, 1936. 



160 



SMELTING 



and Mr. Norman shows that, at least theoretically, it is possible, by 
burning copper concentrates in an oxygen-enriched atmosphere, to 
generate sufficient heat to fuse the roasted product to slag and matte. 
These calculations were based on published analyses of the concen- 
trates which serve as smelter feed at three large smelters Noranda, 
Flin Flon, and Anaconda. 




and AnJtr*on, Am In*t Mm A Met Eng Trans , Vol. 106, p 167, 19SS) 

FIG. 24. Layout of Smelter, Noranda. 

No actual experiments were made, but on the basis of other pub- 
lished work on copper roasting and smelting, Mr. Norman discusses 
some of the problems that would arise in carrying out these ideas in 
practice. The smelting might be done in a down-draft roasting shaft 
(similar to the method used by Laist and Cooper), or it might be 
possible to use a countercurrent system in an up-draft shaft. 

Although such a system would not require extraneous fuel such as 



SMELTER FLOWSHEETS 



161 



coal or fuel oil, it would require a supply of oxygen. Provided that 
a suitable furnace could be designed which would be as satisfactory as 



3 Concentrate Lines 



Smelter Flux 



I Dust O | O Chambers 
Copper Scrap) 




Casting 
Machine 



(Wraith, Am I tut Mm A Met Eng. Trans., Vol 1O6, p. OJ, 1933) 

FIG. 25. Flowsheet, April 1933, Roan Antelope Smelter. 

present equipment, the factor that would be of primary importance 
would be the relative costs of fuel and oxygen. 

SMELTER FLOWSHEETS 

Figures 24 and 25 show the layout or flowsheet of two different 
smelters; these are more or less typical arrangements. 



CHAPTER V 
CONVERTING 

INTRODUCTION 

The operation of matte smelting produces an artificial sulfide or 
matte, and in this chapter we shall consider the problem of treating 
this matte to produce metallic copper. The standard method (con- 
verting) in modern practice is to oxidize the iron and sulfur in a 
converter by blowing air through the molten matte; sulfur goes off in 
the gases as SO 2 and the iron is oxidized and slagged off. The liquid 
metallic copper remains in the converter. Before proceeding to a 
discussion of converting, let us consider the history of the process and 
some of the older methods of treating matte 

Most of the successful methods for converting matte to metallic 
copper have been oxidation processes in which the iron and accom- 
panying sulfur (FeS) are oxidized and removed and subsequently the 
Cu 2 S is oxidized to yield SO 2 and metallic copper. There will not be 
any appreciable oxidation of copper sulfide until practically all of the 
iron sulfide is gone, and this is the fundamental fact upon which the 
success of converting matte depends. 

The Welsh Process. The Welsh process was employed at Swansea, 
Wales, for many years, and until the invention of the converter, this 
method and its modifications constituted the standard treatment for 
matte. It never had any great application in the United States be- 
cause shortly after copper production became important in this country 
the converting process was invented and was speedily adopted. Prior 
to this, however, a great deal of the copper produced in the United 
States was shipped to Wales in the form of matte. 

The Welsh process involved a series of carefully controlled roasting 
and melting operations. The matte was first crushed and given a 
partial roast; this roast was controlled so that the bulk of the iron 
would be oxidized but so that there would not be an excessive formation 
of the higher oxides of iron. The roasted matte was then melted 
with silica or siliceous ore on the hearth of a small reverberatory. The 
iron was removed in a slag (which was usually returned to a blast 
furnace for smelting) leaving in the furnace a high-grade matte 
(white metal) containing 70 per cent copper or better. 

162 



THE WELSH PROCESS 163 

Sometimes a single roast and fusion would produce this rich matte; 
other times it was necessary to employ several partial roasts and 
fusions. Repeated roasts and fusions were helpful in removing arsenic, 
antimony, and other volatile impurities. 

The white metal (essentially Cu 2 S with a small amount of FeS) was 
cast into pigs and treated by the process of blister-roasting. The pigs 
of white metal were melted down on a hearth under an oxidizing flame; 
as the sul fides melted some oxidation took place at the surface and 
the iron sulfide and part of the copper sulfide oxidized: 

2FeS + 3O 2 - 2FeO + 2SO 2 
2Cu 2 S + 30 2 - 2Cu 2 O + 2SO 2 " 

However, the copper oxide formed would immediately react with more 
copper sulfide (roast-reaction) to liberate metallic copper. 

Cu 2 S + 2Cu 2 O - 6Cu + SO 2 

By the time fusion was complete there would be three layers of liquid 
in the furnace metallic copper on the bottom, a layer of molten Cu 2 S 
above this, and a film of slag on top. This slag was formed from the 
small amount of FeO in the matte, silica sand from the furnace bottom, 
and the sand from the casting floor which clung to the pigs of matte. 
The slag was carefully skimmed and returned to the previous smelting 
operation, leaving the layers of copper and matte in the furnace. 

The oxidation of liquid matte by scorification (surface oxidation) 
proceeds very slowly, so no attempt was made to oxidize more of the 
molten Cu 2 S. Instead, the fire was lessened and the charge allowed 
to set or solidify; bubbles of S0 2 evolving from the liquid forced their 
way through the cooling surface so that the crust solidified with numer- 
ous protuberances which gave a large surface for oxidation. After the 
charge had set, it was again heated to fusion and the accompanying 
oxidation formed more copper oxide, which in turn reacted with copper 
gulfide to contribute more copper to the molten layer on the bottom. 
This alternate freezing and melting was continued until practically all 
of the sulfide was gone and the bath had the oily, sea-green surface 
characteristic of metallic copper. The blister-roasting process would 
continue for 24 to 48 hours; at the end the temperature was raised to 
complete the reactions, any remaining slag was skimmed, and the 
copper was tapped into sand molds. Small ariounts of S0 2 gas 
escaping from the metal as it solidified caused blisters to form on the 
surface; hence the name blister copper, a term which is also used for 
the product of the modern converter. 



164 CONVERTING 

Although the converter operation is entirely different in method, as 
we shall see later, the essential chemistry and sequence of operations 
are exactly the same as in the Welsh method. 

The Bottoms Process. If in the blister roasting process the oxida- 
tion is regulated so that about 10 per cent of the contained copper 
is reduced to the metallic state, the metallic copper will remove the 
bulk of the impurities arsenic, antimony, silver, and particularly 
gold as these are more soluble in liquid copper than in liquid matte. 
If such an oxidized charge of matte is tapped from the furnace the 
slabs of metallic copper are found under the first pigs of matte cast, 
and these copper bottoms can then be worked up to recover the con- 
centrated gold. The matte remaining is purer than the original and 
yields a higher grade of copper for the market. 

The universal use of electrolysis for purifying copper has rendered 
obsolete the bottoms process as well as a number of other complex and 
ingenious methods formerly used to recover precious metals from matte. 
In modern converting the gold and other precious metals in the matte 
pass into the copper and are later removed by electrolytic refining. 

Development of Converting. Early attempts were made to oxidize 
matte by the simple proce?>s of blowing air through the molten matte, 
and the successful application of converting or bessemerizing to the 
refining of pig iron lent additional stimulus to these efforts. Between 
1850 and 1855 the converting process was invented in England by 
Sir Henry Bessemer, and it was developed independently in the United 
States by William Kelly at about the same time After 1860 the 
process of converting pig iron to steel rapidly attained great importance. 

The bessemer converter is a pear-shaped refractory lined steel 
vessel with a row of tuyeres in the bottom. It is charged with molten 
pig iron and a blast of cold air is blown through the bath to oxidize the 
metalloids (C, Si, Mn). The first attempts to use the bessemer tech- 
nique for converting copper matte resulted in failures, and a successful 
technique was not developed until the problems peculiar to matte con- 
verting had been worked out. A few of the significant differences 
between matte converting and the converting of pig iron are listed 
below. 

1. Cold air can be blown through liquid pig iron, and instead of 
chilling the metal it actually heats it up because of the heat liberated 
by the oxidation of the iron and impurities. The same thing is true of 
liquid matte but not of liquid copper. The oxidation of metallic 
copper is slow and the amount of heat liberated is small ; consequently 
a blast of cold air will cause liquid copper to freeze rapidly and seal 
up the tuyeres. This meant that the bottom-blown converter could 



DEVELOPMENT OF CONVERTING 165 

not be used, because as soon as liquid copper formed in the bottom 
of the converter it would freeze and clog the tuyeres. Copper con- 
verters are side-blown, and provision is made to allow the metallic cop- 
per to collect below the tuyere level. 

2. About 20 tons of pig iron can be blown in a matter of 18 to 20 
minutes, and the impurities oxidized form a comparatively small 
volume of slag; this is because the pig contains only about 4 or 5 
per cent of impurities to be oxidized and together with the iron loss 
only about 7 to 8 per cent of the metallic bath is removed as gases and 
slag. Matte converting, on the other hand, takes a much longer time 
because there is so much material to oxidize. A 30 per cent copper 
matte, for instance, contains about 40 per cent iron and 30 per cent 
sulfur; thus in 10 tons of matte it is necessary to oxidize 4 tons of 
iron and 3 tons of sulfur, and there remains in the converter at the 
end of the blow only 3 tons of copper or 30 per cent of the original 
weight of matte. 

3. There is comparatively little slag formed in iron converting, and 
usually the silicon and metals oxidized form a siliceous slag (in the 
acid process) which does not have a very severe corrosive action on 
the converter lining In matte converting, the initial blow produces 
large quantities of FeO which rapidly corrodes a siliceous refractory. 
Moreover, if silica is not available to fix iron as a ferrous silicate slag, 
the oxidation continues to form quantities of infusible Fe 3 4 . 

The first successful application of the bessemer process to matte 
converting came in 1880, when Pierre Manhes of Eguilles, France, 
succeeded in producing blister copper from medium-grade matte 
on a commercial scale. Manhes' invention consisted mainly in placing 
the tuyeres above the floor of the converter so that a space was pro- 
vided for the liquid copper to collect. After this demonstration, the 
application of converting was immediate and it soon became the 
standard method for treating copper mattes. 

The early converters were patterned after the bessemer (steel) con- 
verter and were all acid lined with silica or siliceous ore. Thus the 
lining served both as a refractory to protect the converter shell and as 
a flux for the oxidized iron. This resulted in rapid corrosion of the 
lining so that the converter shell had to be relined after every four to 
six blows. 

Development of the basic lined converter was the next step, for it 
was realized that the slow and wasteful method of using the converter 
lining for flux left much to be desired. It was necessary to find a re- 
fractory lining which would resist the action of the basic FeO so that 
relining would not be required at such frequent intervals; the use of 



Punching Rod 




166 CONVERTING 

such a basic lining would of course require the addition of siliceous 
flux to the converter to slag the iron oxide. The first successful use of 
a basic lining was made at the Garfield smelter about 1910; this opera- 
tion was conducted in a Peirce-Smith horizontal converter lined with 
magnesite brick and using silica flux. The superiority of this over the 
acid lining was immediately evident, and soon it was adopted by other 
smelters. Today the basic lined converter is standard, and practically 
all converter linings are made of magnesia either magnesite brick 

or a monolithic lining tamped in po- 
sition. 

Another important invention made 
about this time was the Dibblie (or 
Dyblie) ball-valve arrangement for 
punching the tuyeres. A converter 

FiG.l. Sketch Showing Principle tu ^ re wil1 P lu 8 U P and must be 
of the Dybhe Tuyere Valve. punched at intervals by thrusting a rod 

through it to open it up. The ball- 
valve is located back of the tuyere and a ball is seated in the valve so 
that it directs the air from the tuyere pipe into the tuyere proper. 
When the punching bar is thrust into the hole in back of the ball-valve 
(from the outside of the converter shell) it lifts up the ball and allows 
the bar to be thrust through the tuyere. As soon as the bar is with- 
drawn, the ball drops back into position and seals the hole to prevent 
escape of air. Before the use of this ball-valve, the punching holes 
were closed by wooden plugs or steel caps, which were wasteful of 
air and clumsy to operate. 

CONVERTING OF COPPER MATTE 

Upright Converter. Figure 2 is a diagram of a 20-foot Great Falls 
type converter. This furnace consists of a short cylindrical section 
surmounted by a tapering nose, or head. The converter is mounted 
and swung on two trunnions; one of these carries the ring gear by 
means of which the converter is turned about a horizontal axis, and 
the other is hollow and serves as a wind-box to feed air to the tuyeres. 
The row of tuyeres extends along the back of the converter shell as 
shown in the diagram. These tuyeres are metal pipes extending 
through the shell and the refractory lining. On the outside each 
tuy&re is equipped with punching hole and ball-valve and is connected 
to the air pipe leading from the wind-box. The converter can be 
turned through almost a complete circle about its horizontal axis; it is 
in blowing position as shown in the diagram, and when it is to be 



THE HORIZONTAL CONVERTER 



167 



charged or poured it is tilted forward this raises the tuyere openings 
above the bath so that the air can be shut off. When the converter 
tilted back again, the air is turned on just before the tuyere 



is 



openings are submerged so that the matte cannot run down into the 
tuy&re pipes. Figure 3 shows two Great Falls type converters. The 




^ El. 540 1 .5 

Half Front Elevation ' Half Section 
(Kelly and Laist, Am Inet Mm & Met Eng Trans , Vol 106, p 18, 19SS) 

FIG. 2. Twenty-Foot Converter, Great Falls Type. 

one in the foreground is in the blowing position and the one in the 
rear is tilted forward for pouring. Note that the shape of the nose 
is such that the converter must be tilted well over 90 from its blowing 
position before the liquid can pour out, and it must be tilted almost 
180 before the converter can be completely emptied. 

Upright converters have been made in different sizes, but probably 
the two most common are the 20-foot and 12-foot converters, outer 
diameter of shell (see Fig. 2). One disadvantage of the upright con- 
verter is the fact that the tuyeres are not evenly submerged in the bathj 
in all positions, and this results in an uneven distribution of the blast. / 

The Horizontal Converter. This type of converter is more widely 
used at present than the -Great Falls, or upright, converter. Figures 4, 
5, 6, and 7, show views of Peirce-Smith (" barrel ") converters. These 
have cylindrical steel shells and two steel riding rings resting on sets 
of rollers ;*the weight of the converter is carried on these rollers and 
they permit the converter to be turned about its horizontal axis. Next 
to one of the riding rings is the ring gear which turns the converter. 



168 



CONVERTING 



Air enters through a flexible connection on one side (Fig. 4) of the 
converter, and the tuyeres are connected to the air pipe by short lengths 
of flexible hose or steel pipe. Tuyeres are equipped with Dyblie valves 
and punching holes. 




(Courtesy Anaconda Copper Mining Company) 
FIG. 3. Great Falls Type Converters. 

The opening in the converter for charging and pouring is shown in 
Figure 7. The position of this in relation to the tuyeres may be 
judged by noting that the air pipe which serves the tuyeres is just 
visible at the top of the converter. 

Size of Peirce-Smith converters is commonly given by the outer 
dimensions in feet diameter and length. The two most common 
sizes are 13- by 30-foot and 10- by 26-foot converters. It is a feature 
of the Peirce-Smith converter that the tuyeres are always evenly sub- 
merged regardless of the position of the converter. 

Converter Linings. The old acid linings had a very short life because 
they served to flux the iron oxide, and in most operations a lining 
would only last long enough to produce about 10 tons of copper. Under 
these conditions it was impossible to use the relatively expensive 
silica brick for lining, and the lining was usually made from crushed 



CONVERTER LININGS 



169 



silica or siliceous ore with enough clay to serve as a binder. This 
mixture was rammed into place in the shell and carefully dried until 
the converter was ready for the first charge of matte. The thickness 
of the lining ranged from 12 inches to more than 60 inches (5 feet) ; 







FIG. 4. 13' X 35' Peirce-Smith Converter. 

obviously it was desirable to make the lining as thick as possible so 
that it would last longer. Some linings were so thick that the capacity 
of the converter near the end of a campaign was more than twice that 
of a freshly lined converter. 

A well-designed basic lining will probably last from 200 to 300 
times as long as an acid lining on matte of the same grade, and this 
means that the basic lining warrants greater care and expense in its 
construction. The most common type of converter lining consists of 
magnesite brick set in a single course and backed by grouting of 
ground magnesite and binder such as sodium silirate. The lining may 
range from 9 to 30 inches in thickness, and commonly the lining is 
made thicker in those parts of the converter where corrosion is most 
severe. A new lining is usually dried by warming it with a wood 



170 CONVERTING 

fire in the converter and then treating it with a small amount of matte 
until all the cracks Jretween the bTricks have been filled and the 
lining has a smooth unbroken surface. 

Some linings are coated with a " magnetite lining " by blowing sorqe 
low-grade matte with little or no flux. This produces a large amount 
of magnetite and the slag formed coats the lining of the converter and 
protects the underlying brick. Such a slag is almost infusible because 
of its high content of magnetite, and when this coating wears away it 
can be replaced by repeating the same procedure. 

The mouth of the converter must be of ample size to permit easy 
charging and pouring, to allow the gases to escape easily, and to mini- 
mize the formation of crusts which tend to accumulate around the 
converter mouth. On the other hand, the opening must not be larger 
than necessary because this would mean an excessive heat loss. 

Monolithic magnesia linings have been employed at the United 
Verde smelter 1 with considerable success. The principal component 
of the lining mix is periclase, which is a grain magnesia made by burn- 
ing high-grade magnesite at very high temperatures without fluxing 
impurities (such as are used in making magnesia brick to sinter the 
grains into a compact mass). This periclase will contain from 88 to 
92 per cent MgO. 

The periclase is mixed with water, clay, sulfuric acid and occasionally 
molasses to serve as a binder. The acid is used principally to control 
the pH of the mix, as this has a pronounced effect on the plasticity 
of the clay. The mix is tamped into position and carefully dried by 
means of a wood fire and an oil flame, and the converter is then 
ready for the first charge. Results have shown that the monolithic 
linings last longer than the 15-inch radial brick linings used previously 
and have brought about an over-all saving of about 30 per cent in 
lining costs. 

One of the facts which led to the use of the monolithic linings was 
the patent litigation because of which the company had used acid 
linings during 1925 and 1926; in this period they had perfected 
mechanical methods for tamping linings and had crews trained in this 
work. Such a set of favorable conditions made the adoption of tamped 
basic linings a relatively simple task. 

Converter Air. The air blast entering the converter must be under 
pressure greater than the static head of the matte column over the 
tuyeres. Air pressure used is usually about 10 to 15 pounds per 

1 Parsons, F. H , Development of Monolithic Tamped Periclase Converter Lin- 
ings at United Verde Copper Company Smelter: Am. Inst. Min. <fe Met. Eng. 
Trans., Vol. 106, p. 153, 1933. 



CONVERTER OPERATION 171 

square inch gage pressure. Quick-acting valves in the air line are 
used to turn the air on and off. 

The amount of air entering the converter will depend upon the air 
pressure and the total tuyere area; we shall have more to say about 
this in the next section. The rate at which air enters the bath deter- 
mines the rate of oxidation and hence the heat production and 
temperature rise in the converter. One factor which limits the blowing 
rate is the fact that with too rapid blowing the converter temperature 
may rise too high. 

The amount of air theoretically required to convert FeS to FeO 
and S0 2 and Cu 2 S to Cu and S0 2 can be calculated from the chemical 
equations. Actual practice shows that the air consumed by the con- 
verter will be about 50 to 60 per cent greater than this. The air 
loss is caused by leakage through the lining and by air lost through the 
tuyeres when they are not submerged in the bath and the air is 
still on; also a certain amount of Fe 3 O 4 may be formed which would 
call for more oxygen than that required to form FeO. Analyses of 
converter gases show that practically no free oxygen escapes from the 
bath during the slag- forming stage (while burning FeS). However, 
when burning Cu 2 S during the blister-forming stage, as much as 20 per 
cent of the oxygen entering the bath may escape without combination. 

Experiments have been made using oxygenated air for blowing 
converters, and although the results were promising, this method has 
not been adopted in practice. 2 

Converter Operation. The actual procedure used varies consider- 
ably, depending upon the grade of matte being treated, size of the con- 
verter used, etc. We shall give here a brief outline of the general 
procedure and later we shall consider in more detail the converter 
practice at several representative smelters. 

The converter blow is divided into two stages (1) the slag-forming 
stage and (2) the blister-forming stage. In the first stage the FeS is 
oxidized and slagged by means of the siliceous flux added, thus: 

2FeS + 3O 2 -> 2FeO + 2S0 2 + 223,980 Cal 
and 

o-FeO + ySi0 2 -*(FeO)-y(Si0 2 ) 

The second stage begins when the FeS is all gone, and Cu 2 S begins 
to oxidize: 

2Cu 2 S + 3O 2 -> 2Cu 2 O + 2<30 2 
Cu 2 S + 2Cu 2 O -> 6Cu + S0 2 

2 Tonakanov, S., Blowing with Oxygenated Air: Eng. and Min. Jour., Vol. 135, 
No. 12, p. 539, 1934. 



172 



CONVERTING 



or, by adding and then dividing by 3 we obtain the net reaction: 
Cu 2 S + O 2 - 2Cu + SO 2 + 51,990 Cal 

A charge of matte is added to the empty converter while the con- 
verter is turned down, so that the tuyeres are above the level of the 

liquid matte, flux is added, 
the air is turned on, and the 
converter turned to the blow- 
ing position. Blowing is con- 
tinued long enough (usually 
about an hour or so) to use 
up the flux added, and then 
the converter is turned and 
the slag poured out. More 
matte and flux are added, and 
the blow continued. These 
partial blows are continued 
until all the iron has oxidized, 
and the converter contains 
white metal, which is essen- 
tially pure Cu 2 S. On ordi- 
nary mattes several matte 
charges must be slagged be- 
fore there is enough w r hite 
metal in the converter to blow 
for blister copper. 

During the second stage 
there is no slag formed, and 
as soon as the sulfur is all oxidized the metal is poured out and the con- 
verter is ready for another charge. In some plants the matte is partly 
blown in one converter and the resultant high-grade matte transferred 
to a second converter for finishing. 

Note that the slag-forming reactions generate more heat than the 
blister-forming reaction. The heat generated during the blister- 
forming stage is not much more than sufficient to supply that lost by 
radiation and convection, and the temperature of the converter does not 
rise very much during the second stage. During the slag-forming stage, 
however, the temperature in the converter rises quite rapidly and it 
is necessary to add cold scrap copper, solid matte skulls, and other 
" dope " to cool the converter. 

The capacity of a converter is rather hard to define and does not 
have much meaning except when applied to a given converter 




FIG. 5. Peirce-Smith Converter Turned out 
of Stack Ready for Charging. 



CONVERTER OPERATION 173 

handling matte of a definite grade. It may be expressed in terms of 
the amount of matte treated per day or the amount of copper produced 
per day. 

Let us make a few approximate calculations to indicate what is 
involved in the question of converter capacity. Assuming that matte 
consists of FeS -f Cu 2 S, we can calculate the amount of air theoret- 
ically required to oxidize a unit weight of each. 

2FeS + 3O 2 -* 2FcO + 2SO 2 

2000 

For 1 ton of FeS we shall require - X 3 X 359 = 12,250 cu ft of O 2 

176 

12 250 

at standard conditions or = 58,300 cu ft of air at standard 

\j.Zi\. 

conditions. For the blister-forming stage 

Cu 2 S + 2 - 2Cu + S0 2 
we shall require per ton of Cu 2 S 
2000 



X 359 = 4490 cu ft of 2 
= 21,400 cu ft of air 



160 
or 

4490 
0.21 

ro OAA 

Since = 2.72, if we assume that the blower delivers a constant 

1 ,*-tUU 

amount of air and that the air loss is the same at all stages, we see that 
it takes 2.72 times as long to blow a unit weight of FeS as to blow a unit 
weight of white metal to blister. 

If we know the volume of free air taken in by the blowers per unit 
of time and know the efficiency of air delivery to the converter, we can 
calculate the amount of oxidation in any given period it is all based 
upon the rate at which oxygen is supplied to the molten matte. Mr. 
John S. Stewart 3 has worked out an analysis of this problem for the 
13- by 30-foot Peirce-Smith converter. He assumes that, on an 
average, the air required for a given matte will be 165 per cent of 
that theoretically calculated for the oxidation of the contained iron 
and sulfur. He also points out the importance of taking into account 
the altitude of the plant; a blower might take in the same volume of 
free air per minute at sea level and at an altitude of 6000 feet, but the 
weight of oxygen delivered per minute would be less in the rarefied at- 

3 Stewart, J. S., An Inquiry into Com crtor Capacity Eng and Min Jour., Vol. 
137, No. 5, p. 224, 1936. 



174 



CONVERTING 



mosphere of the higher altitude, and consequently the converting 
capacity would be smaller. Table 1 is taken from Mr. Stewart's 
paper and illustrates the effect of the grade of matte and the altitude 
on the copper output of a 13- by 30-foot converter. These figures 
check quite closely with those found in practice and are based on the 
assumptions that the compressor takes in 20,000 cubic feet of free 
air per minute and that the converter requires 165 per cent of the 
theoretical air. Note how the capacity in tons of copper produced 
varies with the grade of the matte. 

TABLE 1 

METRIC TONS OF BLISTER COPPER PRODUCED DAILY FROM A 13- BY 30-FOOT 
PEIRCE-SMITH CONVERTER 



Altitude 
(meters) 


Grade of Matte, Per Cent Copper 


10 


20 


30 


40 


50 


60 





21.8 


53 


88 6 


129 8 


178 3 


235 8 


300 


21.4 


52 


87.0 


127 5 


175 3 


231 7 


600 


20 9 


51 


85.3 


125 1 


172 1 


227.7 


900 


20.5 


50 


83.6 


122 8 


169 


223 9 


1200 


20 2 


49 1 


82.2 


120 6 


166 3 


220.2 


1500 


19 8 


48 2 


80.7 


118.5 


163.4 


216.7 


1800 


19.4 


47.3 


79 2 


116.4 


160.6 


213.2 


2100 


19 


46 4 


77 7 


114 3 


157 8 


209 4 


2400 


18 7 


45 6 


76 3 


112 2 


155 


206 2 


2700 


18 4 


44 8 


75.0 


110 4 


152 4 


202 6 


3000 


18 


43 9 


73.6 


108.3 


149.8 


199.6 



Liquid matte is poured into the converter mouth from ladles, and 
cold matte skulls, scrap, etc., are usually dumped into the converter 
mouth from " boats "; these ladles and boats are usually manipulated 
by an overhead crane. Flux may be charged by similar boats, by 
means of a hopper set above the converter, or it may be blown into 
the converter by means of a Garr gun set in the end wall of the con- 
verter (the last method applies only to Peirce-Smith, or horizontal, 
converters). The tuyeres are punched at regular intervals while the 
converter is blowing every half hour or so. 

The temperature of the matte added to the converter will usually 
be about 1000 to 1100C; i.e., from 100 to 300 C of superheat 
above its melting point. The rate of blowing and the addition of 
cold material (flux, scrap, etc.) is usually controlled so that the 
temperature in the converter never rises above 1250 to 1300 C. 
Another thing that governs the permissible blowing rate is the fact 



CONVERTER SLAG 



175 



that air pressures much greater than 14 or 15 pounds per square inch 
may cause too much spitting of liquid particles from the converter 
mouth. 

Of course the bath must never be allowed to chill to the point where 
it freezes in the converter. In most plants some sort of auxiliary 
power is provided so that in the event of power failure (and stoppage 
of the air supply) the converter can be turned down and emptied, or 
at least turned down enough to bring the tuyere openings above the 
bath. 

Converter Slag. The slag formed in a converter is essentially an iron 
silicate; while we often assume in calculations that converter slag 
is a ferrous silicate, a good part of the iron is always present as Fe 3 4 
(magnetite) rather than as FeO. This magnetite is partly dissolved 
in the slag and part of it is present as suspended crystals of magnetite. 

Table 2 lists the analyses of three converter slags; note that these 
are high in copper and that the silica is lower than in reverberatory 
slags. The silica content is kept as low as possible to avoid corrosion 
of the basic converter lining; the iron is present as FeO and Fe 3 4 ; slag 
from the Noranda converter, for example, contains about 17.2 per 
cent magnetite. 

TABLE 2 
ANALYSES OF CONVERTER SLAGS 





Per cent 


Sourpp 






Cu 


Si0 2 


Fe 


A1 2 3 


CaO 


MgO 


Noranda 


4 40 


25 


47.7 


4 8 


5 


0.3 


Flin Flon 6 


1 50 


30 


39 8 


4 4 


1 7 


1.0 


Anaconda 


3 95 


23 3 


47 6 


5 9 


1 


0.0 



a Boggs, W B , and Anderson, J. N , op cit. 
b Ambrose, J H., op. cit 

c Laist, Frederick, and Maguire, H J., Reverberatory Furnace for Treating Converter Slag at 
Anaconda* Mining and Metallurgy (A. I. M. E.), No 157, Sec. 13, January 1920 

Copper in Converter Slags.* The copper content of converter slags 
is so high (1 to 5 per cent) that the slags are not discarded but are 
treated to recover as much of this copper as possible. The contained 
copper is found as suspended prills of matte and metallic copper and 
as chemically combined copper silicate. 

During a converter blow the bath is kept in agitation by the streams 
of air, and this tends to disperse a certain amount ol copper and sulfides 

4 Wartman, F. S., and Boyer, W. T., The Form of Copper m Converter Slags: 
U. S. Bur. Mines, Rept. Inv 2985, January 1930. 



176 



CONVERTING 



in the slag these are mechanically entrained in the slag and are 
carried out with it, as no opportunity is given for settling. The copper 
assay of the converter slag, therefore, depends somewhat on the grade 
of the matte. During the first stages of a blow there is less copper in 
the slags removed than in those slags which are formed from blowing 
enriched matte. Very little chemically combined copper is found in 
slags from the first steps of a blow, but as the matte approaches the 

white metal stage, more of this 
oxidized copper is slagged. 

Probably about 90 per cent 
of the copper in average con- 
verter slag is in the form of 
suspended prills of matte and 
copper; only about 10 per cent 
is in the oxidized form. How- 
ever, this copper cannot be 
effectively removed by simply 
allowing the slag to stand; the 
magnetite dissolved and sus- 
pended in the slag tends to 
hold the copper in the slags, 
and it appears that high mag- 
netite content in slags (rever- 
beratory glags as well as con- 
verter slags) usually means 
high copper assays. 

Generally the molten con- 
verter slags are poured back 
into the reverberatory fur- 
nace, and with low-grade matte the weight of slag charged may 
be as great as the weight of new solid material entering the rever- 
beratory. With blast furnace smelting, the converter slag may be 
allowed to solidify so that it can be charged to the smelting furnace, 
or it may be poured directly into the forehearth of the furnace. In 
former days when smelting charges were more siliceous the converter 
slag constituted a valuable iron-bearing flux, but with the basic charges 
that are common at present the converter slag has lost its value as a 
flux often additional siliceous flux must be added to the reverbera- 
tory to take care of the iron. 

About 1915 research was undertaken at Anaconda to determine the 
feasibility of employing a special reverberatory furnace to treat con- 
verter slags. As a result of this investigation a large coal-fired re- 




Fio. 6. Pouring Converter Slag. 



CONVERTER SLAG 



177 



vcrbcratory was constructed (23 ft 4 in. by 153 ft probably the 
longest revcrberatory ever built) to handle the converter slag This 
furnace was eventually abandoned, and the converter slag was charged 
in the regular smelting reverberatones. However, the practice adopted 
here serves as an excellent illustration of what must be done to con- 
verter slag to effectually recover the entrained copper. 

It was found that in order to clean 5 the.se converter slags the fol- 
lowing items were necessary: 

1. The bulk of the Fe ;{ Oj therein contained must be reduced to FeO. 

2. More silica would have to be added to bring the waste slag up to 
at least 30 per cent Si() 2 . 

3. Iron sulfide would have to be supplied to form sufficient bulk 
of low-grade matte to collect the particles of metallic copper and 
sulfides. 

The converter slag furnace at Anaconda attained these ends by 
smelting siliceous ores and calcine* \M(h the molten converter slag. 
In modern practice the converter slag i* added to the smelting reverbcra- 
tory, but it must be remembered that the same general treatment must 



TVHLK 3 

Qt \NTITY OP MM.NETITE ENTEKIM. \vi> LEAVIVG THE RE VERB ERATO RIBS 

\T \OH\VD\ 









^ 


Fe,0, 


Material 


\\ eight 
(tons) 


content 


(tons) 


of total 
input) 


reduced in 
reverbera- 
tones ( r ^) 


Calcine 


<Vi,703 


10 t 


6,729 


47 




Converter slag 


42,,W) 


17 2 


7,319 


53 




Heverl>eratory slag 


70,330 


2 3 


M>1S 


12 


8R 



be given the converter slag to insure the saving of the bulk of the 
copper. There is a tendency for the liquid converter slag to simply 
spread out on the bath in the reverberatory, and for this reason the 
converter slag is usually poured in near the firing end of the rever- 
beratory so that it will react with the pile< of solid charge the sulfides 
attack and reduce the Fe^O 4 and the converter slag becomes diluted 
with more siliceous material. Table 3 shows the amount of magnetite 6 
passing in and out of the Norandn reverberatories for 1 month. Note 

f> Liust, Frederick, and Mafftnre. II J , op. eit. 

rt Bogg, W. B M and Anderson, J. N. The Noranda Smelter: Am. Inst. Mm 
& Met. En. Trans, Vol. 106, p. 187, 1933. 



178 



CONVERTING 



that 53 per cent of the magnetite comes from the converter slag and 
that 88 per cent of the total magnetite is reduced to FeO in the 
reverberatory. The magnetite content of the reverberatory slag is 
much less than that of the converter slag. 




FIG. 7. Pouring Blister Copper. 

Blister Copper. The metal produced in the converter will ordinarily 
contain 98 per cent or more metallic copper. The remainder consists 
of small amounts of base metal impurities Ni, Co, Fe, Sn, Sb, As, Zn, 
and Pb. The exact amount of each of these depends upon the impuri- 
ties in the matte and whether or not any of these metals are volatilized 
in the converter. If the copper is slightly underblown it will contain a 
small amount of sulfur; if overblown it will contain oxygen in the form 
of dissolved Cu 2 0. 

All the precious metals in the matte will pass into the blister 
copper, where they remain until separated by electrolytic refining. 
Selective converting is a method of blowing white metal to form a small 
amount of blister copper which is then separated from the white metal 
before the remainder is blown. This first fraction of the blister formed 
will contain most of the gold, just as did the copper bottoms in the 



CONVERTER PRACTICE 179 

older Welsh process. This method, however, has not been used for 
many years. 

Converter Practice. We shall now briefly summarize the details of 
converter practice at three copper smelters which treat low-, high-, and 
medium-grade matte respectively. 

Noranda. 1 The Noranda smelter employs four Peirce-Smith con- 
verters two are 12- by 26-foot converters with forty 1%-inch tuyeres; 
the other two are 13 by 30 feet and have forty-four 1%-inch tuyeres. 
The converters are served by overhead cranes, and the cranes and 
converter tilting motors are electrically operated. A storage battery 
set is connected to the converter motors so that in case of sudden failure 
of the converter air supply and the power supply to the motors, the 
converters are automatically turned until the tuyeres are clear. 
About 12 pounds per square inch air pressure is used. 

Three of these converters are usually operating at any one time, 
and they treat large quantities of low-grade matte (19.3 per cent Cu) 
with comparatively small production of blister copper. The converter 
slag contains about 25 per cent Si0 2 . Molten blister copper is trans- 
ferred to the anode furnace, where it is fire refined and cast into anodes. 

Roan Antelope . 8 The Roan Antelope smelter is equipped with two 
12- by 20-foot basic-lined Peirce-Smith converters, each with thirty 
1%-inch tuyeres spaced at 6-mch intervals. One converter has 
sufficient capacity to treat all the matte produced in the reverberatory, 
and the other converter is used as a spare. 

The Roan Antelope matte is practically pure Cu 2 S, so little or no 
converter slag is made, and the process corresponds to the blowing of 
white metal at other plants. It was believed at first that this would 
cause difficulties because most converting operations depend upon the 
FeS in the matte as the principal source of heat, and accordingly certain 
special features were incorporated in these converters. The mouth was 
made as small as practical (4 ft 6 in by 5 ft 4 in), and the magnesite 
brick lining was backed by an insulating layer of fireclay brick. A 
14-inch burner port is situated in one end wall of each converter, and 
this is used to burn powdered coal to keep the vessel hot between 
blows. This port is sealed with clay-mud during the blow. 

Actually, the heat generated during the blow is slightly greater than 
the corresponding heat lost by radiation, convection, and in the con- 
verter gases. Temperature readings show that with an empty con- 
verter at 2230 F (1220 C) and a matte charge at 2010 F (1100 C) 
the temperature at the end of the blow will rise to about 2370 F 

7 Boggs, W. B., and Anderson, J. N., op. cit, p. 171. 

8 Wraith, C. R., op. cit., p. 217. 



180 CONVERTING 

(1300C). This means an increase in temperature of about 2.3 F 
(1.3C) per minute. 

The converters are tilted by air motors, and there is always a sufficient 
supply of air at the proper pressure to revolve the converters to a safe 
position in case of failure of the air supply at the tuyeres. Compressed 
air for converting is supplied by a single-stage compressor which is 
rated to deliver 15,000 cubic feet of free air per minute at 15 pounds 
per square inch pressure. 

The average charge to a converter is 60 tons of matte, and this can 
be blown to blister copper in about 2 hours and 25 minutes; during the 
blow the tuyeres must be punched almost continually. The appear- 
ance of the flame and the copper on the rod determines when the 
blow has been completed. 

As no converter slag was available and it was not possible to provide 
a protective layer of magnetite over the converter lining, it was decided 
to charge some crushed hematite with the matte for this purpose. This 
experiment was a success as far as converter operations were con- 
cerned, but it had to be discontinued because the magnetite in the 
converter reverts formed a blanket in the reverberatory ; neither the 
calcic-silicate slag nor the high-grade matte would take up the mag- 
netite either chemically or mechanically; all the magnetite that entered 
the reverberatory furnace had to be removed eventually with rabbles. 

Table 4 lists the converter data at Roan Antelope. 

TABLE 4 
CONVERTER DATA AT ROAN ANTELOPE 

Charge, tons of matte 60 

Air pressure 15 Ib/sq in 

Air consumed per ton of matte 30,590 cu ft 

Free air consumed per minute 13,700 cu ft 

Minutes per converter blow 134 

Minutes per ton of matte treated 2 23 

Minutes per ton of copper produced 2 . 83 

Blister produced per blow 47 . 5 tons 

Blister produced per converter hour 21.3 tons 
Fuel consumed per day 5 9 tons 

Blister copper, per cent Cu 99 6 

The blister copper is transferred into a cylindrical casting furnace 
fired through a pulverized coal port at one end, and from this furnace 
it is cast into 350-pound cakes for shipping. 

Andes. 9 Four Peirce-Smith converters lined with magnesite brick 

9 Callaway, L. A., and Koepel, F. N , op cit., p. 689. 



OTHER CONVERTING OPERATIONS 181 

are used; these are 12- by 26-foot vessels and have thirty-eight 1%-inch 
tuyeres. Converter air at a pressure of 13 to 15 pounds pressure is 
used, and with matte containing 40 per cent copper each converter can 
make about 90 tons of copper per day. The flux used is regular mine- 
run sulfide ore from the fine crushing plant, and it contains about 70 
per cent silica and 16 per cent alumina. Flux is charged from boats 
handled by the converter cranes. The converters have an emergency 
automatic tilting device operated by auxiliary storage batteries. 

Blister copper is poured into two oil-fired cylindrical 8- by 18-foot 
receiving furnaces lined with 9 inches of magnesite brick. From these 
furnaces the copper is cast into blister cakes or anodes. 

OTHER CONVERTING OPERATIONS 

Converting of Lead Mattes. The copper which finds its way into 
lead smelters is usually recovered in the form of a lead-copper-iron 
matte, and this matte is generally treated in copper converters. Most 
of the lead, and also zinc, in these mattes is recovered as an oxide fume 
carried off in the converter gases, and recovered by filtering the gases 
in bag houses. 

Lead and zinc in the metallic form are volatile enough so that they 
can be expelled from liquid slags or mattes at high temperature. 
Their boiling points are 1613 C for lead and 907 C for zinc, but there 
will be volatilization below these temperatures when the opposing 
metal vapor pressure is less than 1 atmosphere. Zinc and lead are 
removed from lead blast furnace slags by blowing powdered coal 
through the molten slag; this reduces the metals to the elemental state 
and they are volatilized and escape from the bath. As soon as the 
metal vapor strikes the air above the bath the metal oxidizes to form 
small particles of oxides which arc carried out by the gas stream and 
recovered in bag houses. Zinc is more readily volatilized than lead 
because of its lower boiling point. 

A similar process permits the recovery of lead and zinc in converters. 
The lead and zinc are reduced by means of the metallic copper or 
ferrous oxide formed in the converter, volatilized and reoxidized, and 
escape in the form of dense clouds of white fume in the converter gases. 
Unless the proper conditions are maintained, however, a high per- 
centage of these metals may be oxidized and fixed in the slag. 

The smelter at Tooele, Utah, 10 has had much experience in con- 

10 Kuchs, 0. M , Lead-Matte Converting at Tooele Am. Inst. Min. & Met. Eng. 
Trans., Vol. 49, p. 579, 1915. 



182 CONVERTING 

verting lead mattes. The composition of such a typical lead matte 
would be: 



Pb 15.0 % 



f o 



Cu 9 05% 
Zn 5.4 % 



S 23.0% 
Fe 37.9% 
Ag 20.3oz/ton 



The original method of treating this matte was to give it a prelim- 
inary blow without siliceous flux to eliminate almost completely the 
lead and zinc. In order to completely eliminate these volatile metals it 
was necessary to overblow the bath, and at the end of the blow the con- 
verter contained a heavy iron slag (FeO and Fe 3 4 ), metallic copper, 
and cuprous oxide. This mixture was then transferred to another 
converter containing a regular charge of copper matte; here the 
iron slag was fluxed, copper oxide was reduced to copper by re- 
action with sulfides, and the blow was carried forward in the regular 
manner. 

Practice at Tooele 11 has been changed since about 1927. Now the 
lead matte is blown with siliceous flux, and the resulting slag is 
skimmed. The matte is then blown to white metal, which is combined 
with white metal from copper matte and blown to blister copper. In 
this method most of the zinc goes into the converter slag instead of 
being volatilized as ZnO as was the earlier practice. 

Converting of Nickel Matte. Nickel and nickel-copper mattes are 
obtained by smelting in the same way as copper mattes, and the con- 
verting of these mattes proceeds in the same way as copper converting 
in the same type of converters. There is one important difference, 
however, in the converting of nickel and nickel-copper mattes the 
blowing can be continued only far enough to eliminate the iron sulfide, 
and the resulting nickel or nickel-copper sulfides must be poured from 
the converter and treated by some other method. Nickel oxidizes so 
readily that the matte cannot be blown down to the metal, and the end 
of the blow corresponds to the " white metal " stage in the converting 
of copper matte. The process is simply one of oxidizing and slagging 
the iron sulfide ; the product of the converter is not a metal correspond- 
ing to blister copper, but a bessemer matte of copper and nickel sulfides 
corresponding to white metal. 

The various methods used to separate the copper and nickel sulfides 
in bessemer matte are beyond the scope of this discussion. 12 

n Sackett, B. L., Converting Lead and Copper Matte at Tooele: Am. Inst. Min. 
& Met. Eng. Trans., Vol. 106, p. 132, 1933. 

12 See Canadian Min. Jour., Vol. 58, No. 11, 1937, for a discussion of the methods 
employed by the International Nickel Co. 



DIRECT SMELTING IN CONVERTERS 183 

Stationary Converter. 13 At the smelter of the Messina Develop- 
ment Company, Ltd., at Messina, Transvaal, South Africa, plans were 
made to treat a copper matte without using standard converter equip- 
ment; this was considered advisable because of the small production 
contemplated (about 20 tons of copper per day from high-grade matte 
containing 60 per cent copper). At first the Nicholls- James process 
was used, in which about two-thirds of the matte was roasted and then 
mixed with the unroasted matte in a reaction furnace; the oxide and 
sulfide copper then reacted according to the familiar roast-reduction 
equation to liberate metallic copper. This process did not work satis- 
factorily because the capacity of furnace and roaster was not great 
enough, and it was decided to employ the reverberatory reaction 
furnace as a stationary converter. After a number of experiments, 
the burner wall was removed and was replaced by a tuyere wall backed 
by a steel tuyere plate. Seven 1%-ineh tuyeres pass through the 
tuyere wall 12 inches above the magnesite furnace bottom at the center. 
Quartz flux is added through a charge hole about 3 feet from the 
tuyere wall. Pulverized coal is burned to provide auxiliary heat, and 
about 7 or 8 pounds air pressure is employed. The maximum bath 
depth is 18 inches above the tuyere level. 

Converting proceeds in much the same way as in an ordinary con- 
verter, except that the technique must be modified because the con- 
verter cannot be tilted to expose the tuyere openings before the air 
is cut off. Special oil-cooled tuyere plugs are used to seal the tuyeres 
when not blowing, and this invention has been largely responsible for 
the success of the method. The stationary converter seems to be best 
adapted for small installations where available labor is cheap and 
not highly skillful. 

DIRECT SMELTING IN CONVERTERS 

Many of the copper concentrates produced today are heavy sulfide 
products containing very little gangue; chemically they resemble the 
artificial sulfide or matte from the reverberatory, and the principal 
difference is that the concentrates are solid, cold, and often moist, 
whereas the matte is tapped from the reverberatory in the form of a 
superheated liquid. The converter burns no extraneous fuel (except to 
keep it warm between blows in a few special cases; e.g., at Roan Ante- 
lope) and is not suited for the smelting of cold concentrates. If, how- 
ever, the sulfide concentrates are added in small amounts to a con- 

18 Knickerbocker, R. G., The Messina Stationary Basic Copper Converter: Am. 
Inst. Min. & Met. Eng. Trans., Vol 106. p. 140, 1933. 



184 CONVERTING 

verier containing hot liquid matte, the sulfides will melt and dissolve 
in the matte and will respond to the converting operation in the same 
way as the sulfides of the matte itself. Naturally the individual sulfide 
additions must be small enough to avoid freezing the matte in the 
converter. 

Practice at Tennessee Copper Company. One of the outstanding 
examples of the use of the converter for smelting copper concentrates 
is the practice at the plant of the Tennessee Copper Company near 
Ducktown, Tennessee. All the copper concentrates treated at this 
plant are smelted directly in a converter. A blast furnace operating in 
1937 on mine-run ore (Chapter IV) produces a 12 per cent copper matte 
and the concentrates are smelted 14 in conjunction with the converting 
of this low-grade matte. 

The converter used is a 12- by 25-foot Peirce-Smith converter which 
is operated continuously during the year except for about two weeks, 
when it is shut down for relming. During this period a Great Falls 
converter, which is kept as a standby, is used. A blast pressure of 10 
to 14 pounds per square inch is used, and a constant-volume regulator 
is used to hold the amount of air to a maximum of 12,000 cubic feet 
per minute this is for the benefit of the acid plant which utilizes the 
converter gases. 

The converter is lined with unburned magnesite brick ; a strip 5 feet 
wide along the tuyeres and running the full length of the converter is 
lined with 20-inch brick, and the rest of the lining is 13 inches thick. 
On starting after relining, a magnetite lining is blown on top of the 
brick by blowing matte until the magnetite layer freezes on the brick. 
The tuyere lining lasts approximately a year, the rest of the lining 
indefinitely. 

Concentrates to be smelted in the converter have the following 
analysis: Cu, 20 per cent; Fe, 35 per cent; S, 35 per cent; Si0 2 , 3.5 per 
cent; Zn, 2 per cent. 

The cycle of operations is as follows: Each morning the copper made 
during the preceding 24 hours is poured, 20 tons of matte is added, 
and the converter is blown for about 20 minutes. Then four 4-ton 
charges of concentrates are added at 20-minute intervals, and before 
all the concentrates have melted, 5 tons of quartz flux is added. 
Blowing is now continued for about 2 hours, and at the end of this 
time the slag is removed. This operation is repeated six or seven 
times during the 24 hours of the day and night, and white metal steadily 
accumulates in the converter until the following morning, when it is 

14 Tennessee Copper Works toward Maximum Economy: Eng. and Min. Jour., 
Vol 138, No 10, p. 40, 1937. 



PRACTICE AT TENNESSEE COPPER COMPANY 185 

blown to blister and poured. Flux is charged from a charging boat, 
and a Garr gun is used to charge the concentrates. 

The first cycles call for a ratio of 16 tons of concentrate to 20 
tons of matte, but after the white metal has built up in the con- 
verter, less concentrate is charged in each cycle. Normally the 
converter handles 120 tons of matte per day, 80 tons of concentrates, 
20 tons of flue dust, and 19 tons of scrap. This means that 0.666 ton of 
concentrate is smelted for every ton of matte converted, under these 
particular conditions; as high as 1 ton of concentrate per ton of matte 
has been treated at this plant. 15 

The following tabulation 16 gives the data for a campaign in 1930-31 ; 
note that the copper assay of the concentrates was lower than that re- 
ported in 1937. 

TABLE 5 

TENNESSEE COPPER COMPANY, CONVERTER CAMPAIGN, JUNE 15, 1930, TO 

AUGUST 23, 1931 

Blowing time (converter days ) 349 . 90 days 

Number of blows 427 

Matte charged 38,335 tons 

Concentrate charged 28,960 tons 

Concentrate per ton of matte 76 ton 

Flux used 11,845 tons 

Copper assay of matte 13 67% 

Copper assay of concentrate 17 13% 

Flux per ton of matte 31 ton 
Flux per ton of copper 1 28 tons 

Flux per ton of iron 37 ton 

Blister copper made 9,246 tons 

Total charge per converter day 235 70 tons 

Time to blow 1 ton of copper 54 50 minutes 

Blast pressure 12 80 Ib/sq ft 

Air used per minute 12,500 cu ft 

Air per ton copper 681,330 cu ft 

Air per ton iron 199,040 cu ft 

Converter slag per day 166 tons 

Of the 166 tons of converter slag, 46.1 tons was in the form of slag 
skulls and was resmelted in the blast furnace. The remaining 119.9 
tons of molten converter slag was poured into the blast furnace settler. 

Entering the blast furnace settler was approximately 3 parts of 
converter slag assaying 1.34 per cent copper and " parts of furnace 

15 Beavers, G. E., Smelting Copper Concentrates in a Converter: Am. Inst. Mm. 
& Met. Eng. Trans., Vol. 106, p. 149, 1933. 
16 Idem, p. 149. 



186 CONVERTING 

slag containing 0.27 per cent copper. The combined slag overflown 
the blast furnace settler assayed 0.35 per cent copper, which meai 
that about 60 per cent of the copper in the converter slag was recover* 
by pouring the slag in the settler. 

TABLE 6 
TENNESSEE COPPER COMPANY, ANALYSIS OF CONVERTER SLAG, IN PER CENT 



Cu 1.34 

Fe 55.3 

8 2.0 



A1 2 O 3 0.70 
SiOo 19 90 

CaO 50 



Smelting Copper-Zinc Ores. 17 Research at Clarkdale on the dire 
smelting of United Verde ore in converters has led to many interestir 
results. The ore in question is a heavy pyrite copper-zinc ore whi( 
is microcrystallme and can not be dissected into copper and zinc coi 
centrates by ore dressing methods, as in the case of Tennessee ar 
Flin Flon ores. 

The experiments carried on at Clarkdale demonstrated that such z 
ore can be smelted in converters to recover copper as blister copper ar 
also to recover a good deal of the zinc and other volatiles (lead, ca( 
mium, arsenic, and antimony) from the converter fumes. The invest 
gation involved the use of many techniques not commonly used wil 
converters, e.g., (1) blowing reducing gases through the tuyeres to n 
duce cuprous oxide and magnetite, (2) use of a second row of tuyer< 
above the bath to burn excess reducing gases issuing from the bat 
and (3) use of preheated air for blowing. Using these various mod 
fications it has been found possible to treat heavy sulfide ores, wil 
the following possibilities as outlined by the authors. 

1. It eliminates the necessity of roasters and roverberatory or bla 
furnaces and smelts coarse pvntie zinc-copper ore direct 

2. No fuel is needed for the smelting and waste heat is available in su 
ficient supply to afford all necessary power for the process and to prehe; 
the air blast 

3. When no fuel is used and molten copper is not recirculated the zii 
vaporization is 50 to 60 per cent in the form of a high-grade zinc oxide fun 
containing most of the other volatile elements of the ore, but copper recovei 
is likely to be low. 

4. By using enough fuel to reduce magnetite to ferrous oxide, copper ca 
be separated with good efficiency. 

^Ralston, 0. C, Fowler, M G, and Kuzell, C. R., Recovering Zinc fro 
Copper Smelter Products: Eng. and Min. Jour., Vol. 136, No. 4, p. 167, 1935. 



SMELTING COPPER-ZINC ORES 187 

5. No flux is necessary and a molten iron oxide product is made which 
may be further refined by gaseous or liquid reducing agents to give a product 
suitable for production of high-grade iron, and simultaneously all remaining 
zinc can be vaporized and recovered as an oxide fume. 

6 Gases can be segregated containing no air and about 14 per cent S0 2 as 
raw material for the production of brimstone. 



CHAPTER VI 

FIRE REFINING 

INTRODUCTION 

Crude Copper. The copper produced by the pyrometallurgical meth- 
ods we have discussed so far is usually too impure for direct use and 
must be refined to produce commercial grades of copper. The impur- 
ities in crude copper fall into two classes (1) base metals and non- 
metals which must be removed because of their harmful effect on the 
properties of the metal, and (2) precious metals which have sufficient 
value to pay for their separation As a rule the precious metals (prin- 
cipally silver and gold) have no deleterious effects on the properties of 
copper, and they may even be beneficial. They are usually present in 
small percentages 100 ounces of silver per ton of crude copper repre- 
sents only about % of 1 per cent. Crude copper may be blister copper 
produced from matte, black copper from the reduction smelting of 
oxidized ores, or crude copper produced by smelting the native copper 
of the Lake Superior district most of this last variety is fire refined 
to produce commercial Lake copper. Crude copper containing precious 
metals is often called copper bullion. Table 1 gives the analyses of 
some representative crude coppers. 

These analyses show that crude copper will contain from 96.5 to more 
than 99 per cent copper and that the percentage of impurities varies 
quite widely. The gold and silver content, in particular, shows great 
variations Noranda blister contains 3.10 ounces ($10850) of gold 
per ton, Tennessee blister contains only 0.05 ounce ($1.75), and the 
crude Calumet and Hecla copper from the melting furnace contains 
practically none. Nkana copper contains 0.0068 per cent bismuth in 
addition to the impurities shown in the table. The amount and nature 
of the impurities determine the type of refining to be used. 

Refinery Location. As crude copper contains approximately 99 per 
cent copper, the shipping weight which could be saved by refining it 
is negligible, and it is often desirable to ship crude copper from the 
smelter to refineries in more favorable localities. Most of the big 
copper smelters are located as close to the mines as possible, and the 
old practice of shipping ores, concentrates, and mattes to distant 
smelters has been discontinued except in a few special places. Many 

188 



ANALYSES OF CRUDE COPPER 



189 



W 

CM 

6 



CQ 



fe 

o 

CO 

w 

rfi 
> 
^ 
< 

< 



+ 


i 


1 


i 


8 


10 

8 


8 




s 

rH 
O 








o 


o 





O 


O 










8 


8 


00 


i 


CO 




rH 




8 
8 




rH 


o 


o 


o 





O 


o 


O 


$ 


* 


8 


1 


3 


3 


10 

8 





O 

3 










o 


O 








"- 1 


o 


a rs 

N ^ 


a 





o 


1 


1 


1 









o 


O 


o 


O 


o 


O 




"*"* 


3 
















CO 


+ 1 


o 


3 


3 


3 


8 


o 










o 


O 


o 


O 


O 


o 







? 


o 


05 

8 


I 


CO 

o 


CO 
rH 

o 


i 




1 
















o 


o 







50 ^ 


I 











i 


i 




i 







c 



















w ? 





GO 

rH 


rH 


i 





I 


3 






o 






















3 1 


o 


CO 
CO 


3 


8 


iO 

o 


o 

rH 




b 





o 


O 





O 


o 


CO 




o 


aJ 


o 

QC 


O 


8 


g 





8 




<N 


N 


10 


CC 


"t 


8 


00 


rH 




N 


3 1 O 


s 


Oi 


8 


?, 


a 


a 


r ^ 


CO 
CO 


Ui^ 


i 


GO 

C5 


00 

C5 











g 


O5 





















1 








\ 






w 




J 





o 









d 









c 

1 

M 


P 

o 

g 


> 


J. 


1 


I 


Calumel 


Nkana c 



190 FIRE REFINING 

ore deposits are in remote places where the problem of supplying the 
necessary power, fuel, and water is considerable ; however, the mining 
operation itself requires considerable power, and all things considered 
it is better to locate the reduction plant (smelter) as close as possible 
to the ore deposit. This applies, of course, to deposits which are 




FIG. 1. Cakes of Blister Copper. 

large enough to warrant the construction of a smelter; for smaller 
mines it is more expedient to install a concentrator and ship the con- 
centrate to a large smelter. 

Many refineries are located at a considerable distance from the 
source of the crude copper, and there are few smelters which are 
equipped to do a complete job of refining. In the United States, for 
example, there are several refineries located on the Eastern seaboard 
which treat crude copper from the Southwest, Africa, and South 
America. Belgium, France, and England also have large copper re- 
fineries for the treatment of crude copper from Africa and elsewhere. 
These locations are favorable because they are close to the large manu- 
facturing centers which use the refined copper and also because they 
are accessible to cheap power. 

Occasionally part of the refining is done at the smelter and part at 
a separate refinery. Anaconda, for example, fire refines its blister and 
casts it into anodes at its smelter in Anaconda, Montana. These 
anodes are then shipped to Great Falls, Montana, for electrolytic re- 
fining. The electrolytic refinery is located at Great Falls because this 
site is adjacent to a large hydroelectric plant on the Missouri River. 

In recent years there has been a greater tendency to build refineries 
nearer to the smelters. Two recent additions are (1) The Montreal 



REFINING FURNACES 191 

East Plant in Quebec, Canada (1931), to refine Noranda and Flin Flon 
copper, and (2) the Nkana plant of the Rhokana Corporation in 
Northern Rhodesia (1935), which is the first electrolytic copper re- 
finery to be erected on the African Continent. 

Refining Methods. There are two methods for refining copper 
fire refining, which we shall discuss in this chapter, and electrolytic 
refining, which will be taken up later. The two methods are closely 
connected, however, because much of the world's copper is treated 
by both. The normal sequence of operations in copper refining is 
(1) fire refining in an anode furnace, from which the copper is cast 
into anodes, (2) electrolytic refining of the anodes to produce cathodes 
of electrolytic copper, and (3) resmelting and further refining of the 
cathodes in a cathode furnace, from which the refined copper is cast 
into wirebars or other commercial shapes. The two fire refining steps 
are similar in principle; the crude blister copper is treated in the anode 
furnace to remove the bulk of the impurities and bring the copper to the 
proper pitch for casting into anodes; the cathodes are fire refined to 
remove sulfur and other impurities taken up by the copper in the 
melting operation, and again to bring the copper to the required pitch. 

Fire refining is an oxidizing operation and is used to remove those 
impurities which are readily oxidized. Electrolytic refining serves to 
remove the impurities which cannot be oxidized ahead of the copper, 
notably (1) the precious metals, (2) bismuth, and (3) small amounts 
of several other elements Ni, Co, Se, Te, As, and Pb. It is gen- 
erally considered that copper should not be refined electrolytically 
unless it contains enough precious metals to pay for the refining or 
unless it contains bismuth, which is not removed in fire refining. This 
classification, however, includes much of the crude copper produced 
throughout the world ; the notable exception in the United States is the 
Lake copper, which ordinarily is not clectrolyzed. 

REFINING FURNACES 

Construction. Fire refining furnaces are called by different names 
depending either upon the nature of the charge or the nature of the 
material being cast thus we have refining furnaces, anode furnaces, 
cathode furnaces, and wirebar furnaces. 

All these arc reverberatory furnaces, but they are small as compared 
with smelting reverberatories. In width they range from 11 to 14 feet 
and are from 26 to 43 feet long as a general average; refining furnaces 
will hold from 120 to 350 tons of molten copper. 

The construction of these furnaces differs in some details from that 
of smelting reverberatories. Many furnaces are constructed prin- 



192 FIRE REFINING 

cipally of siliceous refractories, but there is a tendency to use more 
magnesite and other basic material especially in side walls and 
hearth, which are exposed to the corrosive metal oxides. 

Acid bottoms are commonly made of silica brick, after the practice 
followed in the Michigan copper furnaces; these bottoms are more 
satisfactory than monolithic silica bottoms such as are used in matting 
furnaces. Bottoms may also be constructed of magnesite brick, and 
Noranda 1 has developed a successful basic monolithic bottom made by 
sintering on layers of " Magnafnt," a granular basic refractory con- 
sisting chiefly of magnesia and lime. 

The construction of the bottom is of great importance in refining 
furnaces because of the danger of break-outs of the heavy charge of 
molten copper through the bottom and because the molten copper may 
penetrate cracks and float up parts of the refractory bottom. Brick 
bottoms are laid in the form of a shallow inverted arch, and the brick- 
work is tied in with the side walls. Monolithic bottoms are laid over 
a brick substructure and are sintered on in successive thin layers. All 
these bottoms must be carefully expanded as the furnace heats up to 
avoid cracking. Before using, a new bottom must be seasoned by 
melting a small amount of copper in the furnace and allowing the 
bottom to become saturated with copper. It is essential that refined 
copper be used for seasoning because impure copper would contaminate 
or " poison " subsequent charges. 

Many refinery furnaces have the foundation set on cast iron plates 
which in turn rest on brick pillars; others use some sort of vault under 
the foundation or have pipes embedded in the foundation through 
which air can be circulated. The purpose of these constructions is to 
keep the bottom cool and avoid the break-outs which might occur if the 
refractory bottom were to be softened by prolonged heating at high 
temperatures. 

Refining furnaces are bound together with steel buckstaves and 
tierods, and some employ tension springs on the tierods to keep the 
tension constant as the furnace expands and shrinks on heating and 
cooling. Side walls are made of silica brick, lined with magnesite as 
a rule, and the roof is generally a sprung arch of silica brick. 

These furnaces are provided with openings along the side for 
charging and skimming; these arc closed by means of water-cooled 
metal doors (often copper), or by metal-backed refractory doors which 
can be luted in place with refractory clay. Molten copper is withdrawn 
from the furnace through a tapping slot, which is a slot in the side wall 

1 Boggs, W B and Anderson, J. N., The Anode Department of the Noranda 
Smelter : Am Inst. Min. & Met. Eng. Trans., Vol. 106, p. 329, 1933. 



FUELS 



193 



extending from the lowest point on the hearth to a point above the 
maximum depth of the molten charge. Before the furnace is charged, 
the tapping slot is filled with a stiff mixture of ground refractory and 
water and a number of tapping bars are set in place across the slot 
on the outside of the furnace. These are generally square iron bars 
placed horizontally across the tapping slot, one above another, and 
held in place by lugs on the outside of the furnace. These bars back 
up the refractory in the tapping slot and prevent it from being pushed 
out by the weight of molten copper behind it. The number of tapping 




Section A-A 
(Boggs and Anderson, Am Inst. Mm d Met En>] Trans, Vol 1O6, pp 335-337,1933) 

FKJ 2. Anode Furnace, Noranda. 

bars used will vary from time to time, depending upon whether the 
charge fills the furnace completely or only partially. When tapping 
the furnace, these bars are removed, and the refractory material in the 
slot is broken out a little at a time so that the charge can be withdrawn 
slowly; when the slot has been cleaned out to the bottom, the hearth 
can be completely drained. Figure 2 shows the construction of the 
anode furnace at Noranda, and one view shows a section of the 
furnace through the tapping slot. 

Fuels. Refining furnaces may be fired by pulverized coal, fuel oil, 
or gas, and in general the combustion, burners, etc., resemble those of 



194 



FIRE REFINING 



smelting reverberatories. A discussion of the advantages of various 
fuels used in refining furnaces is given in a paper by Bardwell, 2 of 
the Great Falls, Montana, reduction plant; lump coal on grates, pul- 
verized coal, fuel oil, and natural gas have all been used at the Great 
Falls plant. The data in Table 2 are taken from BardwelFs paper. 



TABLE 2 

COMPARATIVE DATA ON USE OF PULVERIZED COAL, OIL, 
GAS AT GREAT FALLS, MONTANA 



AND 





Pulverized 
Coal 


Oil 


Natural 
Gas 


Calorific value of fuel 


12,:>90 Btu/lb 


19,224 Btu/lb 


983 Btu/cu ft 




11,960 (as purchased) 






Average production tons of 








copper per charge 


235 tons 


233 tons 


233 tons 


Average hourly rate of fuel 








consumption 


2,663 Ib 


188 gal 


32,983 cu ft 


Fuel per ton of copper pro- 








duced 


272 Ib 


19 4 gal 


3,397 cu ft 


Btu per Ib of copper produced 


1,627 


1,441 


1,670 


Average temperature of gases 








leaving furnace 


1940 F (1060 C) 


1940 F 


1940 F 






(1060 C) 


(1060 C) 


Stack temperature 


600 F (315C) 


600 F 


800 F 






(31,5 C) 


(427 C) 


Heat distribution; per cent of 








Btu in fuel. 








Available in furnace 


46 7 


44 7 


41.9 


Absorbed in boilers 


38 9 


40 2 


36 6 


Stack loss 


14 4 


15 1 


21.5 



Refining furnaces are commonly equipped with waste-heat boilers 
to abstract heat from the waste gases. Note that Table 2 shows that 
about 35 to 40 per cent of the total heat in the fuel is absorbed in the 
boilers. Coal ash, which tends to form an insulating blanket on the 
charge, is partly responsible for the higher consumption of coal per 
ton of charge as compared with fuel oil; the higher consumption of gas 
may be accounted for by the fact that the non-luminous gas flame is a 
less efficient radiator than the luminous gas or coal flame. Bardwell, 
however, does not attribute much weight to this last reason and believes 
that the difference is caused by differences in furnace operation or in 
the placing of the burners. It is likely that under the best operating 

2 Bardwell, E S , A Comparison of the Use of Various Fuels in Copper-Refining 
Furnaces Am. In.st. Min. & Met. Eng. Trans., Vol 106, p. 449, 1933. 



THE REFINING PROCESS 195 

conditions for each fuel there would be little variation in thermal 
efficiency. 

THE REFINING PROCESS 

The operation of the fire refining process has changed very little 
since it was employed in the Welsh smelters in connection with the 
older smelting methods; furnaces have been improved and increased 
in size, better charging and casting methods have been developed, and 
the modern refinery has a purer copper to start with; but fundamentally 
the process has not changed. Refining is d batch operation, and one 
charge is refined and cast before more new material is added to the 
furnace. The steps involved in refining are listed below. 

Charging. The manner in which the furnace is charged depends 
upon the form of the crude copper. If liquid blister is being treated 
it is charged by pouring it from ladles through a launder set in the 
side wall of the furnace. This form of charging is usually employed 
when the refining is done in the smelting plant, although the Interna- 
tional Nickel Company 3 has designed a refractory-lined hot metal 
car to haul blister copper a distance of 1% miles from the smelter 
to the anode furnaces. This unit holds 70 tons of molten copper and 
consists of a cylindrical ladle mounted by means of trunnions on 
standard railroad trucks, the ladle is provided with burner ports for 
heating up a cold car or holding a charge under heat. 

Crude copper which is shipped for any considerable distance arrives 
at the refinery in the form of solid cakes or slabs weighing perhaps 
350 pounds each. Scrap charged to the furnace will usually be solid, 
and the cathodes (for cathode furnaces) are flat slabs which weigh 
about 150 pounds. These are placed in the furnace through the side 
doors by means of a charging machine. This machine employs a long 
arm with a fiat paddle, or " peel," on the end; the operator of the 
machine can manipulate the arm in any direction and can turn the 
paddle over to dump its load in the furnace. The solid material is 
loaded on the paddle, and the operator then proceeds to stack it in 
the furnace in such a way as to load in the maximum amount. Figure 3 
shows a picture of one type of charging machine. 

Oxidizing; " Flapping." After the charge has melted down (melted 
" flat ") the oxidation stage begins. Formerly the copper was 
" flapped " or struck with a rabble blade in such a way as to cause 
ripples to travel across the surface; this exposed a greater surface to 

3 Bonarci, Frederic, Transportation of Molten Blister Copper by Rail Am Inst. 
Min. <fc Met. Eng., Tech. Paper 909 (Metals Technology), February 1938 



196 



FIRE REFINING 



the oxidizing atmosphere in the furnace and aided in the oxidation of 
the impurities. The common method used today is to blow compressed 
air into the bath through iron pipes inserted through the side openings 
in the furnace. The oxygen in the air attacks the impurities and 
oxidizes them; the iron pipes gradually burn away and the iron oxide 
enters the slag or skim. 




(Courtesy Anaconda Copper Mining Company) 

FIG. 3. Charging Cathodes into Wirebar Furnace. 

The operation of melting down solid copper slabs takes considerable 
time and makes a longer operation than the refining of liquid blister. 
This is not entirely a disadvantage, however, because the oxidation 
proceeds rather rapidly at the large surface exposed by the melting 
slabs, and a great deal of the oxidation will have already taken place 
by the time the charge has melted flat. This cuts down the time re- 
quired for " flapping " or blowing. Oxidation during melting is more 
rapid than the oxidation caused by blowing air through the molten 
copper. 

As the bath is exposed to oxidizing conditions, the base metal impur- 
ities are oxidized and escape from the metal bath. Sulfur and a 
little of the arsenic and antimony form volatile oxides and are carried 
out by the furnace gases. The other metal oxides together with a 
good deal of copper oxide rise to the surface of the metal bath as a 
viscous slag or " skim "; a little silica is sometimes thrown on the bath 
to slag the metal oxides. The appearance and amount of slag formed 
will depend upon the nature and quantity of the principal impurities 



POLING 197 

in the copper. The slag is skimmed continuously during the oxidizing 
period, and this continues until no more slag forms and liquid Cu 2 
begins to form on the bath. This liquid has a characteristic oily 
appearance, and its presence indicates that the oxidation is complete 
and that the bath is completely saturated with oxygen in the form of 
Cu 2 0. Toward the end of the fining (oxidizing) period, small ladle 
samples are taken and allowed to solidify, and the appearance of the 
surface and fracture of these serves to indicate the condition of the 
bath. As the amount of sulfur in the charge decreases, the bubbles 
(formed by S0 2 ) will disappear from the sample, and \\hen the end of 
the fining period is reached the sample will be brick reH, coarsely crystal- 
line, lusterless, and brittle. This is set copper or copper completely 
saturated with oxygen. 

Cu 2 O exhibits the rather remarkable property of being soluble in 
molten copper, and when the liquid is saturated with oxygen it con- 
tains from 6 to 10 per cent Cu 2 O (0.60 to 90 per cent oxygen) . The 
Cu-Cu 2 alloys freeze with the formation of a eutectic containing 3.45 
per cent Cu 2 0, so that solidified set copper consists of coarse crystals 
of Cu 2 in a matrix of the eutectic. 

When the copper in the furnace has arrived at the set copper stage 
the sulfur and metallic impurities have been largely oxidized; the bath 
is now carefully skimmed and the metal is ready for the next stage, 
which involves the reduction of most of the Cu 2 () back to metallic 
copper. 

Poling. The reduction of the Cu 2 is accomplished by thrusting 
wooden poles into the bath. The heat of the bath causes destructive 
distillation of the wood, and the gases evolved (H 2 0, CO, H 2 , and 
hydrocarbons) stir the bath, and the reducing gases reduce the Cu 2 
to metallic copper. As the poling proceeds, ladle samples of the 
copper are cast and examined, and the appearance of these samples 
indicates the degree of deoxidation of the bath When the copper 
has been poled sufficiently the sample has a metallic luster and 
a rose color, as contrasted with the lusterless, brick-red set copper; the 
crystals are finely radiated and give the surface a silky appearance. 
Copper at this stage is known as tough-pitch copper. __^ 

The poles used are usually green tree trunks 6 to 10 inches in 
diameter; many varieties of wood have been used for this purpose. 
Green wood is preferable because it contains more moisture and gives 
off a greater volume of gas. About six or eight polos will usually suffice 
to treat 100 tons of copper, depending upon how much oxygen is 
present; the oxygen content of set copper depends upon the tempera- 
ture, and the higher the temperature the more Cu 2 that can be 



198 



FIRE REFINING 




(Reproduced by permission from 
Hofman and Hayward, Metal- 
lurgy of Copper, p. 16, McGrau,- 
HiU Book Co., New York, 1924) 

FIG. \. The Cu-Cu 2 O 
Equilibrium Diagram. 



dissolved in the metal (Fig. 4). During the last stages of poling 
(often throughout the entire operation) the bath is covered with a 
layer of charcoal, wood, or low-sulfur coke to prevent reoxidation of 

the copper and to permit the operator to 
hold the copper at the proper pitch during 
casting. 

Tough-pitch copper is not completely de- 
oxidized but contains about 0.05 per cent 
oxygen (0.45 per cent Cu 2 0), and a micro- 
scopic examination of tough-pitch copper 
shows the Cu-Cu 2 O eutectic at the boundaries 
of the large copper grains (Fig. 6). This 
copper has a flat set; i.e., when cast it will 
have a flat surface, it is tough and malleable 
and can be readily rolled, drawn into wire, etc. 
Most commercial metal is tough -pitch copper. 
Overpoling means carrying the copper past the tough-pitch stage 
by removing too much of the oxygen, and this gives the copper un- 
desirable properties. There are always some impurities remaining 
in refined copper because it is impossible to skim the slag perfectly; 
these will ordinarily be in the form of oxidized particles mechanically 
entrained in the copper, and as such are relatively harmless. If the 
reduction is carried too far, these impurities will be reduced to the 
metallic state and will alloy with the copper; in this form they may 
have a pronounced effect on electrical conductivity and other properties. 
Overpoled copper is porous and brittle and does not have the proper 
pitch for casting. This is probably due to the dissolution of reducing 
gases (CO, H 2 ) in the molten copper, for these gases are quite soluble 
if there is no oxygen in the bath to convert them to the relatively 
insoluble H 2 and C0 2 . 

The theoretical explanations of the phenomena encountered in fire 
refining of copper are still inadequate, and many factors are involved 
which are yet to be completely investigated solubility of various 
gases in molten and solid copper; the exact reactions which take place 
and the equilibria involved; effect of "traces" of impurities; etc. A 
review of some of these investigations has been given by Ellis. 4 Some 
of the observed facts for which complete explanations are lacking are 
as follows: 

1. Oxidation and poling will produce a tough, ductile copper con- 
taining in the neighborhood of 0.05 per cent oxygen. This tough-pitch 

4 Ellis, O. W , A Review of Work on Gases in Copper: Am. Inst. Min. & Met. 
Eng. Trans., Vol. 106, p. 487, 1933. 



POLING 199 

copper is the ordinary commercial " copper." By pitch is meant the 
physical properties of copper which regulate its casting and fabricating 
properties thus a copper with the correct pitch can be cast in open 




(Courtesy United States Metals Refining Company) 

FIG. 5. Partly Oxidized Copper Showing Crystals of Copper in a Matrix of the 

Cu-Cu 2 O Eutectic. 

Taken from the " set " (oxidized) surface of a tough-pitch wirebar. The left side of the picture is the 
surface of the wirebar, and the oxygen content diminishes toward the interior. The region to the left 
of the arrow consists almost entirely of eutectic and will contain about 0.38 per cent oxygen. The 
region to the right of the arrow will average about 0.25 per cent oxygen. 

molds to give a dense, sound bar or slab with a flat surface or set ; the 
metal will be tough and can be readily rolled or drawn. If the pitch 
is not correct, however, the metal will be porous and brittle, and the 
surface will not be flat. Although the oxygen content of the copper 
may be the principal factor in determining its pitch, there is evidence 
that it is not the only factor. 

2. Completely deoxidized copper can be produced which is as dense, 
tough, and ductile as tough-pitch copper or more so. The production 
of such copper requires a special technique, however, and it cannot 
be produced by overpoling in the refining furnace. Later we shall 
consider some of the methods used for making this oxygen-free copper. 

3, Slight overpoling can be remedied by " flapping " the bath to 



200 



FIRE REFINING 



bring about slight reoxidation. If the overpoling is decided, however, 
the bath must be reoxidized to set copper and poled again, if the copper 
is to be held at the proper pitch during the casting. 

4. The pitch and set of the copper are affected by the pouring 
temperature and the mold temperature. 




FIG. 6. Photomicrograph of Electrolytic Tough-Pitch Copper " As Cast." 
Shows the Cu-CujO eutectic at the grain boundaries of the copper crystals. 

Composition of Fire-Refined Copper. Table 3 gives the analyses of 
a number of copper anodes copper which has been fire refined but 
has not been electrolyzed. These can be compared with the analyses of 
crude copper as given in Table 1. The anode copper a in Table 3 
represents the average limits of impurities in anodes produced at the 
Nichols refinery in treating a number of crude coppers the analyses 
of some of these are those marked a in Table 1. 

As a rule these fire-refined anodes will contain 99.2 to 99.6 per cent 
copper with not more than 0.4 to 0.8 per cent total impurities. The 
oxygen content of many anodes is higher than that of tough-pitch 
copper which is to be cast into wirebars or other finished products 
(such as would be produced from a cathode refining furnace). The 
gold and silver remain in these anodes as well as small percentages 
of most of the other impurities. Electrolytic refining of these anodes 



COMPOSITION OF FIRE-REFINED COPPER 



201 



<o 

o 

W 55 

>J <J 










CO 




8 fcS 







8 

















o ^ 


2S 


i 


i 1 


1 




O O 


o 







_ 




g 


co cD 




4- ^ 


o 5^ 


CM 


QO 00 

T ^ C5 




$ 


o o 


o 











o 


to o 




GO ^ 




8 


2 Q 3 

o o o 











O O 








Q 


CO 




g 




o 


I . 1 












o o 




6_ 




3 


CO 




-f t^ 


8 fi 





' 


I 


% 


O 














a; 


CO CO 




^ 


g 2 


2 


S 






O 




O 










CO 




S 


S 8 


2 


8 : i 


3 















CD y^O 

< tN 


CO ^ 

O i i 


i 


' 


1 










O 


o 


,| 




$ 


S 8 5 


8 


3, 


O CM 


00 


^ o o 


o 


*i 


O 00 


co 

CO 
1 1 


5 CM 
CO CO 





^^ 


8S 


S 


O5 O *O 


CO 


eg 


! 


S 


S 





g 




c 


1 


w 


1 


3s 


3 


; "0 W fl 







o 


c 


_j5 2 ^ 


-*-3 














o 


Q 


o 2 3 


2 




55 


52; 


S ^ w 


O 



; g-2 



^ <; 



* 



202 FIRE REFINING 

will produce copper containing less than 0.06 per cent total impurities 
as a rule. 

" Refined " and " pure " copper are relative terms ; there is no such 
thing as absolutely pure copper, and the object of the various refining 
methods is to produce a commercial metal which meets certain specifica- 
tions and tolerances set up by the trade. In another chapter we shall 
consider some of these standards, but for the present we shall be mainly 
concerned with the technique and practice of refinery operations. 

REFINERY PRACTICE 

In order to illustrate the application of fire refining methods we 
shall give brief descriptions of some representative fire refining opera- 
tions; the first two operations (Calumet and Hecla, and British Copper 
Refiners Ltd.) produce commercial copper directly from crude copper; 
the third is an example of blister refining to produce anodes (anode 
furnaces) and the fire refining of cathodes (cathode furnaces). 

Calumet and Hecla. 5 ' 6 The furnaces used in refining copper at the 
Calumet and Hecla plant resemble those used for melting of the 
native copper concentrate (Chapter IV) except that jackets are used 
instead of tierods for binding the ends of the refining furnaces. These 
furnaces will hold from 250 to 450 tons of molten copper depending 
on whether the sides and bottom are new or very much worn. 

The charge consists principally of molten copper from the melting 
furnace, and it flows directly from the melting furnace to the refining 
furnace through a launder. Copper oxide precipitate from the leach- 
ing plant is treated in the refining furnace also. A blanket of rich 
concentrates is spread over the furnace bottom to protect the acid 
refractory, and the copper oxide is charged on top of this; after this 
solid material has been partly melted the molten copper from the 
melting furnace is added. 

As soon as the molten copper is in the furnace, air pipes are inserted 
in the bath and oxidation begins; and the charge is completely oxidized 
by the time the solid material is completely molten. The copper oxide 
in the precipitate aids in this oxidation. The principal function of 
the oxidation is to remove iron and sulfur, and silica is thrown on top 
of the bath to slag the iron oxide as fast as it forms. 

5 Lovell, E. R., and Kenny, H. C , Present Smelting Practice at Calumet and 
Hecla: Mining Cong. Jour., p. 67, October 1931 ; Eddy, C. T , Arsenic Elimination 
in the Reverberatory Refining of Native Copper: Am. Inst. Mm. & Met. Eng. 
Trans., Vol. 96, p. 104, 1931. 

6 Hillenbrand, W. J , Poull, R. K., and Kenny, H C., Removal of Arsenic and 
Antimony from Copper by Furnace-Refining Methods: Idem, Vol. 106, p. 483, 1933. 



BRITISH COPPER REFINERS, LTD. 203 

After the oxidation is complete and the iron slag has been skimmed, 
green hardwood poles are inserted and the bath poled to about 0.04 to 
0.05 per cent oxygen this tough-pitch copper is then ready for casting. 
The slag is skimmed, granulated in water, and then dried and mixed 
with coal screenings to be charged back into the melting furnace. 

The native copper in the Lake deposits is extremely pure, and 
hand-picked pieces of the metallic copper display electrical properties 
and purity superior to the best electrolytic copper. In addition to the 
copper, however, the deposits contain heavy arsenic minerals and 
occasionally native silver. These follow the copper into the con- 
centrate and become alloyed with it in the melting funace. To remove 
the arsenic a special treatment is used. 

After the iron slag has been skimmed, powdered soda ash is blown 
into the bath while the copper is still oxidized but before poling begins. 
Most of the arsenic reacts with the soda ash thus: 

2As + 50u 2 O - lOCu + As 2 O 5 
As 2 5 + 3Na 2 CO 3 -> 2Na 3 As0 4 + 3C0 2 

The sodium carbonate melts at the furnace temperature and forms 
a liquid slag on the surface of the bath ; the sodium arsenate is soluble 
in this slag and is thus removed. Any antimony present is removed 
by a similar set of reactions. The soda ash slag is very fluid and 
extremely corrosive to the furnace brickwork so it must be skimmed as 
soon as possible. After this slag is removed the poling and casting 
proceed in their normal order often the poling is started even before 
the soda ash slag is skimmed. 

Copper treated by the soda ash process is sold as " Prime C. and H." 
copper and meets all the specifications required of electrolytic copper. 
For some architectural and other uses, arsenic is a desirable constituent 
because it imparts greater resistance to corrosion; so two other brands, 
" Natural C. and H." and " CL " brands, are made in which the 
soda ash treatment is not used Typical analyses of the three grades 
are given in Table 4. 

In addition to these brands, some metal is sold on the basts of silver 
content. High-silver Calumet and Hecla copper is used for special 
purposes which require the metal to maintain its strength at high 
temperatures. 

British Copper Refiners, Ltd. 7 The refinery of British Copper Re- 
finers, Ltd., is located at the city of Prescot, about 8 miles northeast 

7 Aldrich, C. H , The Fire Refinery of British Copper Refiners, Ltd.: Am. Inst. 
Min. & Met. Eng. Trans , Vol. 106, p. 467, 1933. 



204 



FIRE REFINING 



TABLE 4 
ANALYSES OF LAKE COPPER 





Prime 
C. and H. 


Natural 
C. and H 


CL 


Copper + silver 


99 9500 


99 9100 


99 6545 


Arsenic 


0025 


0445 a 


2000 6 


Iron 


0.0025 


0025 


0.0025 


Nickel 


0.0015 


0.0015 


0.0015 


Sulfur 


0015 


0015 


0015 


Oxygen 


0420 


0400 


0.0400 



Arsenic content ranges from 02 to 00 per cent 
6 Arsenic content ranges from 06 to 50 per cent. 

of Liverpool, England. The refinery treats Roan Antelope blister 
from Africa. 

The refining furnace has a capacity of 200 tons of copper and is 
fired with pulverized coal. The furnace is set on 4-foot concrete piers 
to permit cooling of the hearth; these piers are capped with %-inch 
steel plates spread with a thin layer of graphite. On top of the piers 
rest the 2-inch thick ribbed cast iron plates which carry the furnace 
bottom. These plates are separated by 1-inch spaces to provide for 
expansion, and the graphite prevents binding to the metal caps on 
the piers. 

The furnace bottom is shaped by a layer of concrete composed of 
ganister and cement tamped in place; sheets of % 6 -inch steel are laid 
over this concrete and the silica-brick bottom laid on top of these steel 
sheets. The working bottom consists of two 12-inch inverted arches 
of silica brick laid dry with broken joints, and any cracks are filled 
with hot pulverized silica. The brick in these arches is keyed in such a 
way that if any of the brick in the upper layer wears thin and floats up, 
the bottom layer can be depended on to hold the charge. Side walls 
are made of magnesite brick to about 6 inches above the metal line. 
The roof is a sprung arch of 15-inch silica brick laid dry and with 
broken joints. 

In charging the furnace, four or five wheelbarrows of sand are first 
thrown on the hearth, and then a thin layer of wire scrap is added to 
serve as a cushion. The cakes of blister copper are then stacked in 
the furnace with the charging machine, and melting is started. When 
the charge is melted (" off bottom ") the slag is skimmed, and the air 
pipes inserted. Blowing is continued and the slag is skimmed at 
intervals until the " say ladle " sample indicates an oxygen content 
of 0.90 per cent; then the slag is skimmed clean and a fairly thick 



ONTARIO REFINING COMPANY 205 

layer of low-sulfur coke is spread over the bath. Any rejected wire- 
bars or refined scrap which is to be remelted are added at this 
point. 

Poling is carried out in the usual manner, using a good grade of green 
hardwood poles. When the oxygen content has reached 0.03 to 0.04 
per cent the poling is stopped and the charge cast into wirebars. The 
bars produced are equal in appearance to the best electrolytic wirebars 
and show an appreciable margin over the requirements of standard 
specifications for electrolytic copper. 

The slag produced amounts to 2 or 3 per cent of the weight of the 
charge. This is skimmed into steel boxes, and after solidification it 
is crushed and shipped to an associated plant for further treatment. 

Ontario Refining Company. 8 - 9 ' 10 The Ontario Refining Company 
treats principally the blister copper from the Copper Cliff smelter of 
the International Nickel Company; it maintains both anode and 
cathode furnaces. 

Anode Furnace Practice. The plant has three anode furnaces with 
space provided for a fourth. These are rated at 300 tons capacity, but 
charges up to 350 tons are handled regularly. The inverted arch of 
the furnace bottom is made of concrete, which rests on cast iron plates 
supported on 4-foot concrete piers; the working bottom consists of 
two layers of 12-inch silica brick. The walls below the metal line are 
15-inch magnesite brick, and above the metal line they are 15-inch fire- 
clay brick. The roof is a sprung arch of 20-inch silica brick; the 
skewbacks are of water-cooled steel. The verb is made of chrome 
brick, and chrome brick dividers are used throughout the furnace be- 
tween magnesite and acid brick. There are four charging bays in 
each furnace; two of these are closed by water-cooled doors and the 
other two have doors of clay brick in a steel frame. Doors are 
operated by hydraulic lifts. The life of the various refractories in the 
furnace is approximately as follows: roof, 100 to 125 charges; magne- 
site side walls, 200 charges; clay-brick side walls, 100 charges; and 
bottoms, 3 to 5 years, by patching the top layer when needed. 

Prior to 1936 all the incoming blister reached the refinery in the form 
of 460-pound cakes. These were cast from two 150-ton holding 

8 Benard, Frederic, Electrolytic Copper Refinery of Ontario Refining Company: 
Am. Inst. Mm. & Met. Eng. Trans., Vol 106. p 369, 1933 

9 Benard, Frederic, Transportation of Molten Blister Copper by Rail from 
Smelter to Refinery: Am. Inst Min. & Met. Eng , Tech Paper 909 (Metals 
Technology), February 1938. 

10 Benard, Frederic, An Investigation into Anode-Furnace Refining of High- 
Nickel Blister Copper: Am. Inst. Min. & Met. Eng., Tech. Paper 910 (Metals 
Technology), February 1938. 



206 



FIRE REFINING 



furnaces at the smelter and 




reaches the refinery furnace 



hauled to the refinery. The practice em- 
ployed when charging solid blister is 
outlined below. 

The first charge to the furnace 
was a layer of anode scrap spread 
over the floor to protect it from the 
impact of the copper pigs. The blis- 
ter cakes or pigs were then charged 
by means of a crane-operated casting 
^ machine. Other miscellaneous feed 
to the anode furnace included more 
& anode scrap, silver-refinery slag, 
3 tank-house and storage-building 
sweeps, bosh scale, metallics from 
slags, etc. About 325 tons could be 
charged in 2 hours. 

After charging, the charge was 
melted, and oxidized to about 0.60 
per cent oxygen; then the slag was 
skimmed, the bath covered with a 
layer of coke and the charge poled 
until the oxygen content was low 
enough to give a flat set. Green 
poles of white and yellow birch were 
used, and 5 to 6 tons of poles would 
be required per charge. The anodes 
cast from the furnace contained 
about 99 per cent copper with vary- 
ing amounts of gold and silver; the 
& chief base-metal impurity was nickel, 
j> which would average about 0.45 per 
g cent. The slag produced averaged 
fe about 2 per cent of the weight of the 
charge, and this was crushed and re- 
turned to the smelter. Pulverized 
coal was used for firing, and the coal 
burned amounted to 11 to 12 per 
cent of the weight of the charge. 

Practice since 1936 has been es- 
sentially the same as that outlined 
above except that the blister copper 
in the liquid form. In 1936 a specially 



hC 
p 

I 



8 



be 
a 



O 



g 



ONTARIO REFINING COMPANY 



207 



designed car was developed to haul molten blister from the smelter 
1% miles away. A standard-gage track was laid in the anode-charging 
aisle so that the car could be spotted directly along the side of the 
furnace and its contents poured into the furnace through a launder. 
The car holds about 70 tons of metal, which means that 4 to 5 carloads 
are required to provide a furnace charge (300 to 350 tons). 

Table 5 shows the time required for a complete furnace cycle, using 
solid blister and liquid metal. 

TABLE 5 

COMPARISON OF FURNACE CYCLE FOR BLISTER CAKE AND 
HOT METAL, ONTARIO REFINING COMPANY 



Process 
(300-ton charge) 


Time (hours) 


Blister Cake 


Liquid Metal 


Charging 
Melting and skimming 
Flapping 
Poling 
Casting 

Complete cycle 
Percentage of slag 


2 
14 


15 
5 

3 

7 


23^ 

2% 


30 

5% 



The holding furnaces previously used at the smelter were coal-fired 
reverberatories ; blister from the converters was charged directly 
into these furnaces, slag was skimmed, and the charge was poled 
sufficiently to give a flat set to the blister cakes. Under the new system 
the blister copper goes directly from the converters to the transfer 
car and from there to the anode furnace. The entrained converter 
slag which was formerly removed in the holding furnaces now enters 
the anode furnace, and this accounts for the fact that the slag pro- 
duction is more than doubled. About one-third of the blister copper 
coming from the smelter is Or ford process copper (copper obtained from 
the working of copper-nickel bessemer matte), and this metal is so 
highly oxidized that ordinarily no " flapping " or oxidizing is necessary 
by the time the copper is charged and skimmed it will contain about 
0.9 per cent oxygen and is ready for poling. 

The total time for a furnace cycle has been increased by about 
6 l /2 hours over the old practice; this is because each anode furnace 
actually serves as a holding furnace for about 15 hours while enough 
blister copper is being produced to fill it. During the filling period 



208 FIRE REFINING 

sufficient coal is burned to keep the charge at approximately 2250 F 
(1232C), and slag skimming begins as soon as the charge level is 
high enough. As there must always be furnace room available for the 
molten blister, the furnaces are operated on a staggered schedule so 
that there is always one furnace being filled. These refining furnaces 
are equipped with two tapholes, and under the old practice both were 
used for casting; the new schedule makes it more desirable to use 
only one taphole, and accordingly the time required for casting a 
charge has been doubled. 

Blister copper reaching the refinery will contain from 0.50 to 1.50 
per cent nickel, and the anode furnace treatment removes about 15 
to 20 per cent of this. The rest of the nickel is removed in the electro- 
lytic tanks, and it appears that the oxygen content of the anodes (as 
regulated by poling) has its effect on the dissolution of nickel from the 
anodes. Low-oxygen anodes (0.03 to 04 per cent oxygen) seem to 
give the best results because most of the nickel dissolves in the electro- 
lyte (which is desirable from the refinery standpoint). When the oxy- 
gen content of the anode is higher (0.10 to 30 per cent) less of the 
nickel dissolves, and more of it is found in the electrolytic slimes. 
The reason for this behavior is not clear, and this is one more example 
of the interrelation of the oxygen content, impurity content, " set," 
and " pitch " of copper for which a complete explanation is lacking. 

Cathode, or Wirebar Furnaces. The furnaces, charging aisle, waste 
heat boilers, casting equipment, etc , in the wirebar furnace building 
practically duplicate the anode furnace equipment. The cathode 
furnaces are also rated at 300 tons capacity, but the normal charge is 
about 320 tons. This is less than the anode furnace charge because 
a part of the refined copper produced is sold as sheared cathodes and 
hence is not remelted in the cathode furnace. 

The cathodes from the electrolytic tanks are slabs of highly purified 
metallic copper about 3 feet square and l / 2 inch thick; each cathode 
weighs about 240 pounds. These may be sheared into smaller sections 
for remelting in the brass trade, to be used as anodes for copper 
plating, etc., but most of them are remelted in the cathode furnace to 
be cast into wirebars to be rolled and drawn into wire; the rough 
surface of cathodes makes them unsuited for direct mechanical fabrica- 
tion. When these cathodes are melted down in a fuel-fired furnace 
the copper absorbs small amounts of impurities (principally sulfur), 
and to remove these and bring the copper to the proper pitch the metal 
must be carried through the complete cycle of melting, flapping, and 
poling. 

At Ontario the cathodes are stacked in the furnace by means of a 



ONTARIO REFINING COMPANY 209 

crane-operated charging machine. The bath is oxidized by blowing 
compressed air through the metal until the oxygen content is about 
0.90 per cent; slag is then skimmed and the bath poled down to 0.028 
per cent oxygen to give tough-pitch copper. Coal consumption is 
about 12 to 13 per cent of the total charge, and the slag formed 
amounts to 1.75 per cent. A complete cycle requires about 24 hours, 
broken down as follows: 

Charging, l| to 2 hours 

TV/T IA j i/\ * 11 u f Flat, 6i hours 

Melting period, 10 to 11 hours { > a ' , , 

^ [ Afloat, 4 hours 

f Flapping, 4 hours 

Oxidizing, poling, and casting, 11 to 12 hours \ Poling, 3 hours 

I Casting, 5 hours 

The temperature of the copper at various stages is given as follows: 

F C 

" Off bottom " or afloat 2150 (1177) 

Flapping, early stages 2140 (1171) 

Flapping, before coking 2175 (1191) 

Poling, 1 hour after coking 2125 (1163) 

Poling, 2 hours after coking 2100 (1149) 

Ready to cast, about 3 hours after coking 2070 (1132) 

Casting 2060 (1127) 

The cast copper (wirebars and other shapes) will contain 99.96 per 
cent copper, 003 per cent oxygen, and a total of about 001 per cent 
of all other impurities. Complete analyses of this and other electro- 
lytic coppers will be given in Chapter VIII. 

Summary. The examples quoted above illustrate the application of 
fire refining methods although, of course, they do not illustrate all the 
different modifications of furnace practice. The following general 
observations may be made. 

1. Fire refining may be used alone to produce a marketable grade of 
copper, and when it is so used the product is often equal to the best 
grade of electrolytically refined metal. 

2. Copper containing precious metals in any quantity will be re- 
fined electrolytically, as this is the only practical method for their 
separation. The electrolytic process also removes the base metal 
impurities associated with the copper. 

3. Bismuth is one metal which cannot be removed satisfactorily by 
fire methods. The Nkana refinery, for example, was built to treat 
copper which contains too much bismuth to make a satisfactory fire- 



210 FIRE REFINING 

refined product; the Nkana blister does not carry sufficient gold and 
silver to warrant electrolytic treatment for their recovery alone. 

4. Blister copper is generally fire-refined before casting into anodes; 
this removes some of the impurities and thus makes the electrolytic 
refining easier. In some cases, e.g., Mount Lyell, Tasmania, 11 a high- 
grade blister may be cast directly into anodes. Fire-refined anodes 
will usually contain about 99.2 per cent copper; these will be produced 
from crude copper which may contain as little as 96.0 per cent copper. 

5. Copper for anodes is usually poled only enough to give a flat set, 
and the oxygen content may be as high as 0.30 per cent. Finished 
shapes such as wirebars and cakes for mechanical fabrication must be 
poled to the tough-pitch stage with an oxygen content of about 0.03 
per cent; this applies whether the treatment is all fire refining or the 
melting of electrolytic cathodes. 

CASTING OF COPPER 

The casting of copper into suitable shapes involves a number of 
considerations temperature of the metal, " pitch " and " set " of the 
metal, and the type of casting equipment used. The old method of 
casting copper was to dip the copper from the furnace with hand 
ladles and pour it into molds; modern furnaces are too large for this 
method of casting, and today practically all copper is cast by mechan- 
ical methods. 

The care which must be exercised in making copper castings depends 
upon the purpose for which the finished object is intended. In casting 
blister copper, for example, it is only necessary that the cakes be flat 
enough to stack properly, and irregularities and blow-holes on the 
surface are not of great importance. Wirebars and cakes for rolling, 
on the other hand, must have smoothly finished surfaces so that when 
rolled there will not be any defects in the wire or sheet. 

Molten copper has a strong tendency to absorb oxygen from the 
atmosphere, and therefore it must flow from the furnace to the mold 
through the shortest possible distance. The flow of metal from the 
furnace is usually a steady stream, and this flows into a tilting spoon or 
ladle from which it flows into the mold. The interruption of the 
flow which is necessary when replacing a filled mold by an empty one 
is brought about by raising the lip of the pouring ladle for a short 
time. Meanwhile, of course, the copper continues to flow from the 
furnace into the pouring ladle. The lip of the pouring ladle should be 
as close to the mold as possible. 

11 Murray, R. M, Electrolytic Copper Refining at Mount Lyell, Tasmania: 
Am. Inst. Min. & Met. Eng. Trans , Vol 106, p. 408, 1933. 



CASTING BLISTER COPPER 



211 



Casting Blister Copper. Blister copper (and other forms of crude 
copper) is usually cast in the form of large cakes (Fig. 1) which will 
weigh from 350 to 450 pounds. The blister is usually stored in a 
holding furnace as it comes from the converters and is cast from this 




(Courtesy Traylor Engineering and Manufacturing Company) 
Fia. 8. Twenty-Two-Foot Casting Wheel. 

Two molds are in place on the wheel. The pouring ladle and its control mechanism can be seen in 
the foreground. 

furnaces resembling refining furnaces; others are cylindrical furnaces 
which are nothing more than Peirce-Smith converters without the 
tuyeres and equipped with burners to keep the charge hot. The 
cylindrical furnaces are convenient for casting because they can be 
tilted easily for pouring. 

When blister is to be cast it is often taken from the converter while 
slightly underblown and still contains about 0.1 per cent sulfur this 
gives a sounder casting. If the blister copper is overblown it may be 
necessary to pole the blister copper enough to bring the oxygen content 
down and give the copper a flat set. Some slag may be skimmed 
from the holding furnace if necessary. 

The blister copper molds are carried on a mechanically driven 
horizontal wheel or on a " straight-line " casting machine by means 
of which the molds are successively brought under the pouring lip of 
the casting ladle. The molds are usually made of copper and are 



212 



FIRE REFINING 



cast as needed by means of a master mold. When cylindrical holding 
furnaces are used the blister may be poured directly into the molds 
without the use of a pouring spoon or ladle. After the filled molds 
pass under the pouring lip the copper begins to solidify; and as soon 
as it is frozen, sprays of water are played on the cakes to cool them. 




(Courtesy Anaconda Copper Mining Company) 
FIG. 9. Casting Copper Anodes in a Straight-Line Casting Machine. 

Casting Anodes. The liquid copper from which anodes are cast is 
usually furnace-refined copper although blister may be cast directly. 
Anode molds are commonly made of refined copper and may be made 
from a cast iron master mold or cast in sand. The molds are carried 
on casting wheels or on straight-line machines (Fig. 9), and the copper 
enters the mold from a pouring spoon into which it flows from the 
furnace. Flow of the copper from the refining furnace is regulated 
by gradually cutting the burned clay out of the tapping notch as the 
level of the metal drops. 

It is essential that the pouring spout be controlled by a mechanism 
which gives a smooth motion so that there is no sudden surge of metal 
into the mold to cause splashing. The casting machine must be 
driven so that it accelerates and decelerates smoothly, because any 



CASTING ANODES 



213 



sudden jerks of the molds would mar the surface of the solidifying metal. 
Figure 9 shows the casting of anodes in a straight-line machine, and 
Figure 10 shows some finished anodes. All anodes have the same gen- 
eral shape a flat rectangular slab with supporting lugs at the top 
but they may vary somewhat in size and weight. Anodes such as those 




FIG. 10. Copper Anodes. 

shown (for multiple refining) will be approximately 3 feet square and 
about 2 inches thick; smaller anodes will weigh about 500 pounds 
and larger ones up to 750 pounds. The cast lugs serve to support the 
anode in the electrolytic tank and one of them makes contact with the 
busbar which carries the electric current. Some anodes are cast with 
a "Baltimore groove" in one lug a special notch which gives a 
better contact between anode and cathode bar. 

The following description applies to casting of anodes on a casting 
wheel; the procedure with a straight-line machine would be essentially 
the same except that the molds would move in a straight line instead 
of a circle. 

For the first round or two of the machine only about % inch of 
copper may be poured in each mold to dry and warm the molds; these 
are scrapped and remelted. After this each mold 



214 FIRE REFINING 

is filled to the proper depth ; then the pouring ladle is tilted back and 
the wheel turned until the next mold is under the pouring lip. After a 
filled mold has moved through % to % of the circumference, the 
surface will have frozen over, and from here on water sprays are played 
on the surface to aid the cooling. After the mold has moved far enough 
so that the anode is completely solidified it is lifted from the mold by an 
automatic device and transferred to a water bosh or tank through which 
a large volume of cold water is circulating. This quickly chills the 
anodes down to room temperature, and then they are taken from the 
bosh for inspection. Defective anodes are scrapped, and the good 
anodes are racked for transport to the tank house. 

The empty mold is sprayed with a wash (usually a slurry of 
powdered silica or bone ash in water) to prevent the copper sticking 
to the mold and continues on its journey around the wheel until it again 
comes under the pouring lip. 

Casting of Refined Copper. Refined copper for the market is usually 
cast on a casting wheel. In many respects the process resembles that 
used in casting anodes. These are large horizontal wheels carrying 
the molds on the circumference. The central wheel or turntable rests 
on rollers running in a circular track and is driven by an electric 
motor. The two principal types of casting wheels are the Walker and 
Clark machines. On the Walker wheel the molds are placed with 
their long axes along the circumference of the wheel; on the Clark 
machine the molds are set parallel to the radial arms of the machine. 
Figure 8 shows a casting machine set up in the manufacturer's 
plant; two of the molds are in place, and in the foreground can be 
seen the pouring ladle and the control platform from which the ladle 
is tilted. 

In general the casting procedure follows that used in casting anodes 
" warmers " are made in the cold molds, the molds are sprayed with 
a wash (which dries immediately on the hot mold) to prevent sticking, 
sprays are used for cooling, and the final cooling takes place in a water 
bosh. Usually the molds are turned over automatically to discharge 
the piece into the bosh. 

Great care is taken to insure a finished shape having no defects; 
the temperature of both mold and liquid metal must be controlled, and 
of course, the copper must have the correct " pitch." Finished shapes 
are carefully inspected before being shipped. 

Where a variety of shapes are made on the same wheel, a large 
supply of molds must be kept on hand ; each type of mold may require 
a different type of pouring ladle. Figure 11, for example, shows a 
pouring ladle with five spouts pouring wirebars into a five-bar mold. 



VERTICAL CASTING OF COPPER 



215 




FIG. 11. Pouring Wirebars. 

Vertical Casting of Copper. 1 2 The top or " set " surface of copper 
cast in an open mold is somewhat wrinkled, and it often has a high 
oxygen content and some porosity. Copper cast in " flat " molds 
(Fig. 9) with a short vertical dimension will have this " set " surface 
on its largest face ; the faces in contact with the mold are comparatively 
smooth. When copper wirebars or cakes are rolled dcr.m into rods or 

12 Strom, B. H., Vertical Casting of Copper at Carteret: Eng. and Min. Jour 
Vol. 136, No. 2, p. 59, 1936. 



216 



FIRE REFINING 



sheets, the set surface may result in surface imperfections on the 
finished object. 

In order to diminish the amount of this set surface, the copper 
refinery of the United States Metals Refining Company at Carteret, 
New Jersey, has developed a method of " vertical " casting in which the 
mold for cakes is placed with its long dimension vertical, and the set 
surface then appears on the end rather than on the face of the cake. 
Vertical casting has been used even for the relatively long, thin 
wirebars. Other refineries throughout the world have been licensed 
to use the vertical casting method, and it is used in casting shapes for 
certain particular specifications. 

It is often necessary to " scalp " or machine the set surface off wire- 
bars or cakes when these are to be used to produce wire or sheet with 
carefully finished surfaces; the use of vertically cast shapes obviates 
the waste and expense of the scalping operation. , 




FIG. 12. Copper Ingots. 

Commercial Shapes of Copper. Following are the shapes in which 
most refined copper appears on the market. These are castings, or 
refinery products; of course much copper is sold in semi-fabricated 
forms such as rods, bars, tubes, plates, and sheets, but these are all 
originally made from refinery shapes such as are listed below. 

Shapes for Remelting. Copper for remelting or alloying is commonly 
sold as ingots (Fig. 12) , cathodes, and " warmer bars." Ingots weigh 



COMMERCIAL SHAPES OF COPPER 217 

about 20 pounds apiece, and ingot bars (essentially two or three ingots 
cast together) weigh about 50 pounds. These ingot bars are notched 
so that they can be easily broken into smaller pieces. Cathodes may 
be sold just as they come from the electrolytic tanks, but usually they 
are first sheared into smaller pieces. 

Copper sold for melting and alloying must meet certain required 
chemical and electrical specifications, but physical defects such as 
imperfect bars, shrink holes, and concave tops are of no consequence. 
The shapes listed below, however, are used primarily for direct fabrica- 




FIG. 13. Copper Ingot Bar. 

tion, and in addition to chemical and electrical specifications, these 
shapes must meet rigid requirements as to freedom from surface defects 
in set and casting, and they must not show more than a certain 
specified variation from standard weight. 

Wirebars. Because of the large amount of copper used for rod and 
wire, the wirebar is one of the most common shapes for refined copper 
some refineries cast their entire output in the form of wirebars. 
These shapes are used for rolling to rod, which may then be drawn 
into wire. They are long rectangular rods (Fig. 14) about 3% to 4 
inches square and tapered at both ends to facilitate rolling; they range 
in length from 38 to 54 inches and in weight from 135 to 300 pounds. 
The cross-section is not perfectly square because the molds are slightly 
tapered so that the bar can be readily removed. Ordinary wirebars 
have a set surface on the largest flat face; when desired, "scalped" 
wirebars or vertically cast bars can be supplied which do not have this 
set surface. 

Cakes and slabs are used principally for rolling to sheet; these are 
approximately rectangular in section and of various sizes and shapes, 
depending upon the product to be rolled. Vertically cast cakes 



218 



FIRE REFINING 



(" wedge cakes ") have the set surface on the end where it is relatively 
harmless. 

Billets are round bars cast on end, and they are principally for 
piercing in the manufacture of seamless tubing; often they are 



'M-^O U.. * ' 






FIG. 14. Copper Wirebars. 

made of deoxidized copper. Some billets are also used for rods and 
other shapes made by the extrusion method. Billets range from 2 
to 10 inches in diameter and from 75 to 750 pounds in weight. 

DEOXIDIZED COPPER AND OXYGEN-FREE COPPER 

Table 6 gives some typical analyses of commercial copper ; note that 
some of these types (tough-pitch coppers) contain about 0.03 to 0.04 
per cent oxygen whereas others, including the cathode copper, contain 
no oxygen. As we have already noted, the presence of this small 



DEOXIDIZED COPPER AND OXYGEN-FREE COPPER 219 

amount of oxygen is essential to get sound castings from the ordinary 
refining furnace. Tough-pitch copper yields castings with a slightly 
crowned, almost flat set, and it freezes without the formation of 
cavities or pipes. 

The removal of this small amount of oxygen produces some important 
changes in the properties of the metal. Oxygen-free copper is ex- 
ceptionally ductile, is easily welded, and can be used where it must 




(Courtesy United Stales Metals Refining Company) 

FIG. 15. O F H C Copper " As Cast." 
Compare with Fig. 6. 

be heated in a reducing atmosphere. If ordinary tough-pitch copper 
is heated in a reducing atmosphere above 400 C the reducing gases 
react with the oxide particles at the grain boundaries (see Fig. 16) and 
thus form cracks which cause the section to become brittle. Thus 
there are certain uses for which copper must be completely free of 
oxygen, and we shall now consider some of the methods of making 
such copper. Copper free from oxygen forms a deep pipe or shrinkage 
cavity on freezing, and this portion must be cropped and discarded or 
eliminated by feeding with molten metal during the solidification 
period. 

The difference between deoxidized and oxygen-free copper lies in 
the method of manufacture; deoxidized copper is tough-pitch copper 
that has been treated with a deoxidizing agent, and oxygen-free copper 



220 



FIRE REFINING 




(Courtesy United States Metals Refining Company) 

Fia. 16. Tough-Pitch Copper, 12-Gage Wire Heated m Hydrogen 

Atmosphere for 1 A Hour at 850 C. 
Note the " attack " on the gram boundaries. 




(Courtesy United State? Metah Refining Company) 

FIG. 17. O F H C 12-Gage Wire Heated in Hydrogen Atmosphere 
for M Hour at 850 C. 

Compare with Fig. 16. 



DEOXIDIZED COPPER AND OXYGEN-FREE COPPER 221 



8'S d 

"5 8 

0> T3 <3 

H > 

O 



8 



8 3 8 8 8 






O O O 

odd 







*O CO CO 

888 

o o d 



888 

000 



a 



o o o 

d o d 



8 8 



rH CO IO 

i ~ 

d 



8 8 



rH rH (N 

888 



O 

d 



co ^ 
o o 



S 8. 

" 3 



CO 

o 



ppe: 



I 



O 



*. 15 I 



a 
& 

o 

I 



O 



^ '03 

3* 
*1 

^1 
^^ 

w 



"& 



El 



vi 



- 
i 






nical 



8 






1 



222 FIRE REFINING 

is made by treating cathodes under such conditions that no oxygen is 
allowed to enter the copper. Note that the cathodes (Table 6) are free 
from oxygen. 

Deoxidized Copper. Deoxidized copper is made by adding a strong 
reducing agent to molten copper. The copper is usually in the 
tough-pitch stage (0.03 to 0.05 per cent oxygen), and it may be taken 
directly from the refining furnace, or it may be obtained by remelting 
solid copper. The deoxidizing agent is added to the ladle or crucible 
containing the liquid metal, and it combines with the residual oxygen. 
The resultant oxide is insoluble in the melt, and it rises to the top; 
some will remain in the metal as mechanically included particles. 

Phosphorus in small amounts is commonly used where the copper 
is not to be used for electrical purposes; it is necessary to add a 
small excess of phosphorus to insure removal of the oxygen, and this 
excess phosphorus alloys with the copper and greatly decreases its 
electrical conductivity (Table 6) . Other deoxidizing agents are silicon, 
calcium, lithium, and calcium boride the last three do not have 
much effect on the electrical conductivity. 

Deoxidized copper forms a deep pipe when it freezes, so usually the 
upper part of the cast billet is cropped and scrapped. 

Oxygen-Free Copper. Oxygen-free copper is made by special 
methods, of which we shall describe two. 

F H C Copper. 13 Commercial copper known as " F H C " 
(oxygen-free, high-conductivity) copper is made by a patented process 
developed by the United States Metals Refining Company at Carteret, 
New Jersey. The plant contains a 75-ton, oil-fired reverberatory 
melting furnace (Fig. 18) which operates continuously; cathodes are 
fed in as fast as the copper is tapped out. The bath is poled con- 
tinuously at the end opposite the burners to keep the oxygen content 
between 0.03 and 0.05 per cent. A constant stream of metal flows from 
the taphole of the furnace at a temperature of 1150 C. 

The molten copper flows through a refractory lined trough under a 
cover of charcoal and enters the deoxidizing unit. This is a specially 
designed refractory-lined vessel filled with high-grade wood charcoal 
over which the stream of molten copper trickles. The residual oxygen 
is reduced thus: 

C + Cu 2 O -* 2Cu + CO 
CO + Cu 2 - 2Cu + C0 2 
C + CO 2 -> 2CO 

The copper leaving the unit is completely deoxidized. 

13 Cone, E. F., Oxygen-Free High-Conductivity Copper: Metals and Alloys, 
Vol. 8, No. 2, p. 33, 1937. 



OXYGEN-FREE COPPER 



223 



The deoxidized copper passes through a special spout and into a 
closed launder which conducts it to the pouring hearth. The entire 
atmosphere in the spout, launder, and pour hearth consists of a special 
charcoal producer gas which contains about 27 per cent CO, 0.50 per 
cent CO 2 , and the balance N 2 ; this gas should be free of H 2 , H 2 0, and 
hydrocarbons. 



Temperature control panel for 
"De Ox" unit and pour hearth 




\^ v ^xc;;:;x\^x^^\\\^^^ 

(Cone, Metals and Alloys, Vol 8, No 2, p. S3, 1937) 

FIG. 18. Equipment Used in the Production of O F H C Copper. 

The pouring hearth is an elongated cylindrical hearth in which a 
large bath of metal is constantly maintained; its purposes are (1) to 
regulate the temperature, and (2) to control the stream of metal during 
pouring. This hearth contains two low-frequency induction heaters, 
and by means of these the temperature is held within a 10 C variation. 
The pouring hearth is rocked mechanically to control the pouring, and 
the copper enters the molds through a spout enclosed in a special hood. 
This hood and the pouring hearth are kept filled with a controlled 
atmosphere of charcoal producer gas, and the bath of metal in the 
pouring hearth is covered with a layer of charcoal. These methods 
illustrate the difficulty of keeping molten copper from absorbing 
oxygen deoxidized copper will pick up as much as 0.01 per cent 
oxygen in flowing from the furnace to the mold in an atmosphere of air. 

The oxygen-free copper is cast into wirebars, billet?, and cakes, all 
of which are cast vertically. These shapes all contain the shrink hole 
or pipe near the top which is characteristic of oxygen-free copper; the 
upper portion of each casting is cropped and scrapped. 



224 FIRE REFINING 

One of the routine tests which is regularly made on samples of the 
finished shapes demonstrates some of the qualities of F H C copper 
as compared with tough-pitch copper. The sample is forged and 
drawn into 0.08-inch wire, and the wire is annealed in a hydrogen 
atmosphere at 850 C for 30 minutes and then quenched. The wire is 
then subjected to a reverse bend test through a 90 angle. F H C 
copper must stand 10 reversals without breaking, and as many as 12 or 
15 are common. Tough-pitch copper subjected to the same treatment 
usually breaks after one reversal. 

Coalesced Cathode Copper. 14 A recent process takes advantage of 
the fact that cathode copper is essentially oxygen-free copper and 
that the oxygen content of tough-pitch copper is caused by the 
exigencies of melting and casting. In the new process the absorption 
of oxygen is avoided because the copper is fabricated without ever 
becoming molten. 

Small particles of cathode copper are first briquetted by compressing 
them at a pressure of 20,000 pounds per square inch. This is performed 
at room temperature and yields a coherent briquette which has a 
density of 80 to 86 per cent that of solid copper. The briquette is 
then heated in a reducing atmosphere consisting of such gases as 
propane gas, nitrogen, steam, and hydrogen at 1600 to 1670 F 
(871 to 910 C). The briquette then passes directly to the ex- 
trusion press through a controlled-atmosphere chamber which prevents 
oxidation. 

In the extrusion press the copper is forced through a die at pressures 
of 30,000 to 53,000 pounds per square inch and emerges as a rod (or 
other shape, depending on the die used) of solid metallic copper. Under 
the high pressures in the extrusion chamber, the copper particles 
coalesce and develop a complete new grain structure; for this it is 
essential that the particle surfaces be clean and unoxidized, and this is 
the reason for the treatment in the deoxidizing atmosphere. The 
copper produced exhibits the same general properties that are charac- 
teristic of oxygen-free copper prepared by other methods. 

One of the necessary requirements for this process is a suitable 
method for obtaining electrolytic copper in the form of small pieces, 
and research has solved this problem by developing methods for pro- 
ducing brittle cathodes. We shall describe these when we discuss the 
electrolysis of copper. 

14 Tyssowski, John, The Coalescence Process for Producing Semifabricated 
Oxygen-Free Copper: Am. Inst. Mm. & Met. Eng., Tech. Paper 1217 (Metals 
Technology), June 1940. 



ELECTRICAL MELTING OF CATHODES 225 

Electrical Melting of Cathodes. Another method for the melting of 
cathode copper is the use of electric furnaces. The following quotation 
is taken from The Mineral Industry. 15 

An interesting innovation of possible far reaching importance is the intro- 
duction of a new method of converting cathode copper into finished shapes. 
The International Nickel Co. has installed an electric arc furnace which is 
fed continuously with cathodes and delivers finished copper to vertical billet 
molds without going through the usual blowing and polmg operations. 
Details of the operation are not available for publication, but it may well 
be that this will prove to be the forerunner of many similar operations. 
It has always seemed incongruous that material with the purity of cathode 
copper should require such a cumbersome treatment as that usually accorded 
it just to put it in a form suitable for use. 

It should be remembered that the real pioneer in the use of an electric 
furnace on a large scale was the United States Metals Refining Co. in their 
early work on F.H.C. copper. They used an induction furnace rather 
than an arc furnace and the relative merits of the two must still be settled. 
It may even be possible that the same results obtained at the International 
Nickel plant may be reached in a reverberatory furnace provided special 
precautions are taken to keep a reducing atmosphere. The important thing 
is that it has been fully demonstrated that first class wire bars can be 
produced without oxidizing and poling in the accepted way. 

15 The Mineral Industry During 1937, Vol. 46, p. 196, McGraw-Hill Book Co., 
New York, 



CHAPTER VII 

SMOKE AND GASES 

INTRODUCTION 

In the previous chapters we have considered roasting, smelting, 
converting, and fire refining, and although we have discussed the solid 
and liquid products of the various furnaces, we have only casually 
mentioned the gaseous products. Perhaps this has been an unconscious 
holdover from early metallurgical practices, when roasting was done 
in heaps and stalls and smelting furnaces were equipped with short 
individual stacks which discharged the waste gases directly into the 
atmosphere. The damage to surrounding vegetation and the " stack 
losses," both in heat and in metallic values, caused by this wasteful 
practice soon led to the development of methods for better handling of 
waste gases. At this point we shall briefly consider the general prob- 
lem of smelter smoke and the methods used in handling smoke and 
gases. 

The importance of this question may be seen from a single example 
A reverberatory smelting furnace treating 800 tons of charge per day 
and burning 112 tons of coal would require about 38 million cubic feet 
of air (1500 tons) and would produce about 40 million cubic feet of 
waste gases (1650 tons) . The heat carried by these waste gases would 
represent the equivalent of 50 to 60 tons of coal, and in addition these 
gases would carry off in suspension perhaps 80 tons of the charge in 
the form of dust and fume. The copper loss in the dust might easily 
amount to 5 or 10 tons a day, and in addition there might be con- 
siderable quantities of gold, silver, arsenic, etc. contained in the 
dust and fume. 

The methods used for handling smoke and gases vary considerably, 
and the practice used at any given smelter will depend upon local 
conditions. In general there are four important facts to be considered 
about any particular smoke : 

1. The amount produced per day. 

2. The nature of the gaseous constituents. 

3. The temperature of the smoke, and its sensible heat content. 

4. The nature and amount of suspended matter carried by the smoke. 
The terms waste gases, flue gases } and smoke are used rather loosely 

226 



COMPOSITION OF SMOKES 227 

and somewhat indiscriminately. Strictly speaking, the term gas or 
gases should be restricted to material which carries no solid or liquid 
matter in suspension ; a smoke is gas carrying a certain amount of sus- 
pended matter, and it is this that renders it visible; all true gases are 
transparent, and the common gases found in metallurgical smokes are 
also colorless. 

Gases produced in pyrometallurgical operations are discharged into 
the atmosphere, but before this is done, it is necessary to 

1. Remove the suspended matter. 

2. Abstract as much of the sensible heat as is practicable. 

3. Remove the S0 2 or dilute with other gases to cut down the con- 
centration of S0 2 . Where it is not possible to remove all the S0 2 , the 
smoke should be discharged into the upper atmosphere by means of a 
tall chimney. 

COMPOSITION OF SMOKES 

Gases. The principal gases found in the smokes from copper smelt- 
ing operations are nitrogen, water vapor, carbon dioxide, carbon mon- 
oxide, oxygen, and sulfur dioxide. None of these gases have any 
commercial value except S0 2 ; when this is present in sufficiently high 
concentration the gases can be used for the manufacture of sulfuric acid 
or sulfur compounds. 

Sulfur dioxide is also the only one of the true gases which is harmful 
to vegetation, and as it may also yield valuable byproducts, it is the 
most important from the standpoint of treatment and disposal of 
waste gases. 

In all the pyrometallurgical operations that we have considered 
(except electric smelting) , air has been used to burn either carbonaceous 
fuel or sulfides, or both, and as air contains 79.0 per cent by volume of 
nitrogen (including about 1.0 per cent argon and other inactive gases) 
which passes unchanged through the reactions, it follows that nitrogen 
will make up a large percentage of all smokes. Copper smelter smokes 
will contain from 73 to 77 per cent nitrogen as a rule. Carbon dioxide 
will be present if carbonaceous fuel is used ; carbon monoxide is seldom 
found except in very small amounts. Water vapor will always be 
present, and the amount depends upon the moisture (if any) in the 
furnace charge, and the amount of hydrogen and moisture in the fuel. 
Let us briefly consider the approximate analyses of the gases produced 
in different operations. 

Reverberatory Matte Smelting. Reverberatory furnaces usually 
operate with a draft of about 0.1 inch water gage, and the combustion 



228 SMOKE AND GASES 

is usually regulated so that a slight excess of air is used. The 
analyses of reverberatory flue gases will be: 





Per Cent 




Per Cent 


N 2 
CO 2 
CO 


72 to 76 
10 to 17 
Oto 0.1 


2 
H 2 O 

S0 2 


5 to 60 
4. Oto 10 
l.Oto 2.0 



The amounts of C0 2 and H 2 will depend upon the fuel used, and 
the amount of moisture on the charge; the S0 2 content will depend 
upon the sulfur elimination from the charge. A large part of the free 
oxygen found in these flue gases may be due to leakage of air through 
charging holes and other openings in the furnace. These gases will 
leave the furnace at a temperature of 1800 to 2300 F (980 to 
1260 C). 

Refining. The gases from reverberatory refining furnaces are 
essentially the products of combustion of the fuel used; firing con- 
ditions resemble those of smelting reverberatories, but the amount of 
gas evolved from the bath is comparatively small. Following are 
some typical gas analyses: 





Per Cent 




Per Cent 


N 2 


73 to 77 


2 


2 to 40 


CO 2 


8 to 15 


H 2 


5 to 16 


CO 





SO 2 


0.1 



The exit gases from refining furnaces will have a temperature of 
1900 to 2000 F (1040 to 1090 C). 

As a general thing, two facts are characteristic of the gases from 
reverberatory furnaces (both smelting and refining furnaces) (1) the 
high temperature of the gases makes it possible to use waste-heat boilers 
and other devices to recover much of their sensible heat content, and 
(2) the S0 2 content is too low to make suitable raw material for 
acid manufacture. 

Converters. The gases from converters consist principally of N 2 , 
S0 2 , and 2 . The oxygen which passes through the bath is largely 
consumed, but some air is drawn into the flue around the converter 
mouth. Converters are blown into hoods (Fig. 4) and there is enough 
space between the converter and hood to permit the ingress of cold 
outside air. This dilution, together with the intermittent operation 
of the converters, makes it difficult to satisfactorily recover the sensible 
heat from converter gases. The S0 2 content of converter gases will 
range from 3.0 to 13.0 per cent. 

Roasters. Roaster gases contain principally nitrogen, oxygen, and 
sulfur dioxide, the amount of S0 2 ranging from 4.0 to 9.0 per cent. The 



FUME 229 

amount of S0 2 in the gases depends upon the amount of sulfur in the 
charge ; in copper roasting the conditions may vary from simple drying 
of high-copper concentrates (with the use of auxiliary fuel) to the 
autogenous roasting of heavy pyritic concentrate, and the S0 2 content 
of the gases will vary accordingly. The gases from blast roasting will 
resemble hearth-roaster gases in composition. Roaster gases are not 
as hot as the gases from smelting and refining furnaces ; they will gen- 
erally leave the roaster at a temperature of 1000 F (540 C) as com- 
pared with 2000 F (1093 C) for the reverberatory flue gases. 

Blast Furnaces. Gases from blast furnaces smelting heavy pyrite 
ores and concentrates contain N 2 , CO 2 , H 2 O, and S0 2 ; the amount 
of C0 2 will range from 2.5 to 4.0 per cent, depending on the amount 
of fuel used, and the S0 2 content w r ill usually be about 6.0 to 7.0 per 
cent. Blast furnace gases are cooled by passing upward through a 
column of charge which abstracts much of the sensible heat. The 
gases from furnaces smelting oxide ores or concentrates (either blast 
furnaces or reverberatories) will be practically free of S0 2 . 

Electric Furnaces. The gases evolved from electric smelting fur- 
naces (Chapter IV) are different from all the other gases considered 
in that no blast of air is required for the furnace, and consequently the 
waste gases are not diluted with such a large volume of nitrogen. If 
the electric furnace is sealed to prevent air leakage, the bulk of the 
waste gases comes from the charge itself, and the furnace gases will 
contain from 10 to 20 per cent SO 2 . This is a decided advantage when 
it is desired to recover the sulfur. There is a possibility that gases 
much richer in S0 2 might also be made in roasting and converting if 
the practice of using oxygen or oxygen-enriched air were to be adopted. 

Dust. The amount of dust carried out in stack gases will depend 
upon the fineness of the particles on the charge, the type of furnace, 
the method of charging, etc. The dust itself may include anything in 
the furnace charge which is fine enough to be carried by the gas current. 

Fume. Fume, as differentiated from dust, refers to material which 
has been volatilized or sublimed, and then condenses when the gases be- 
come cooler. The most important constituents found as fume in cop- 
per smelters are 

1. The lower, volatile oxides of arsenic and antimony As 2 3 and 
Sb 2 3 . 

2. Oxides of other volatile metals e.g. PbO and ZnO. 

3. Condensed water vapor. 

4. Sulfuric acid and sulfates. A certain amount of S0 3 gas is 
formed from the further oxidation of S0 2 , and the higher the S0 2 con- 
tent the greater will be the amount of S0 3 . The S0 3 combines with 



230 SMOKE AND GASES 

water vapor to form droplets of sulfuric acid (H 2 S04 + water), or it 
may combine with certain basic oxides, notably ZnO, to form ZnSO4. 
Smokes which contain free acid are known as acid smokes; basic smokes 
contain an excess of basic oxides, and any free acid is neutralized. 

In practice the dust and fume are often mixed together and collected 
as a single product; usually the collected product is called a dust e.g. 
flue dust, Cottrell dust even though the bulk of the material may be 
a true fume. 

Table 1 gives the chemical analyses of a few smelter dusts taken at 
random. Note the extreme variation in composition and that the 
copper content is usually rather high. 

WASTE-HEAT RECOVERY 

Of the various methods which have been employed for recovering 
waste heat from furnace gases, by far the most important is the use of 
waste-heat boilers on reverberatory smelting furnaces 

Waste-Heat Boilers. 1 ' 7 Most of the discussion and illustrative ex- 
amples given in this section are taken from a symposium on waste-heat 
boiler practice in the United States as published in the Transactions of 
the American Institute of Mining and Metallurgical Engineers. 

Waste-heat boilers are standard equipment on practically all rever- 
beratory furnaces. These boilers are usually set directly in front of 
the furnaces so that the furnace gases strike the boiler tubes as soon 
as they leave the furnace laboratory. The boilers themselves are 
usually of the vertical water-tube type, and the Stirling boiler (Figs. 1 
and 2) is the most widely used Boilers may be arranged in different 
ways each furnace may have a single boiler, or the furnace may 
have two boilers arranged in parallel or in tandem (Figs. 1 and 2). 

Boiler practice does hot greatly differ from that of direct-fired boilers 
in most respects, and we shall not consider such questions as boiler-feed 

1 Bardwell, E. S., Copper-Refinery Waste-Heat Boilers at Great Falls: Am. Inst. 
Mm & Met. Eng Trans., Vol 106, p 225, 1933 

2 Barnard, E A., and Tryon, George, Waste-Heat Boiler Practice at Anaconda 
Idem, p. 230. 

3 Sager, N. W., and Mossman, H. W , Waste-Heat Boiler Practice at Nevada 
Consolidated Copper Corporation: Idem, p 237. 

4 Marston, J. R, \Vaste-Heat Boiler Practice at United Verde: Idem, p. 246. 

5 Honoyman, P. D. I , and Faust, P. A , Waste-Heat Boiler Practice at Miami 
Idem, p. 251. 

6 Rose, J. H., Waste-Heat Boiler Practice at the Magna Copper Company 
Smelter- Idem, p. 255. 

7 Marriott. R. A, Waste-Heat Boiler Practice at the Garfield Smelter: Idem, 
p. 257. 



ANALYSES OF SOME DUSTS 



231 



H 
O 
(C 

& 



< 

H 



W 


*> r-fe 




; 











^3 


rH CM ~* 






























* 


^ CO CM 








a 






u 


O O CM 








o 








CM t>- Ci 





CO 


00 


s 





CO 





<* CO 
rH rH 


5 


CO 
CM 


$ 





8 


CO 


94 

O 


O l^ 00 


CO 


o 


00 


00 
00 


O 


00 


CO 


O) *O O^ 


00 


8 


rH 





oo 


CM 




I- rn b. 






CM 


5 


CM 


CM 




O O *O 






t " 


3 


CO 





s 


rH rH 






























OH 


CO CM 














C 


CM "O 


CM 


CM 










S3 


o co 


^ 


^ 










o 


CO 














cc 


CM Q 














W 


CO C 

rH ^ 














_ 


o r^ CM 


CO 


s 


5 


a 


C5 


$ 


O 


CM Ci rn 

CM rn CO 


N 


CO 


^ 


CO 


!J 


6 




1 1 " 
















"k ^ 1 






























c 
o 


T3 T3 C 










^ 


, 


1 

1 


Q? 

S E S E ^ 
ErS E-S b 

2.8 ^^ | 
x ^ >> ^> i., 

|l |l 1 

II la 1 


6 

0) 
3 
<C 


er Cottrell dust. c 


ir Cottrell dust.* 


Converter dust * 


1 

T3 

! 
1 

1 


rter Cottrell dust 




u & ? * .. 


"S 











8! 


* 


<V u c; M V 
U K ^ 5 

1 1 1 

| 7 I- 1| ^ 1 


i 

1 


s 



S 


1 


d 
s 


a 

8 




1 J Z J 5 S 


a 


,s 


n 


X 


2 







< K 


E 






K 


jz; 


S5 



? 





u 

I 

^j 



232 SMOKE AND GASES 

water and removing scale from the interior of the water tubes, which 
apply to all boiler operations. The operation of waste-heat boilers in- 
volves some additional factors which are discussed below. 

The steam generated in these boilers may be used for many purposes 
around the plant heating, generation of power, etc. It must be re- 
membered, however, that these are essentially waste-heat boilers and 
that the smelting or refining furnace is not primarily a steam-producing 
unit; in refining furnaces, for example, the firing conditions vary con- 
siderably during the different stages of the refining cycle, and conse- 
quently the steam production is not constant. Waste-heat steam 
should not be relied upon to produce a steady and uninterrupted supply 
of steam power, as it would not be practical to operate the furnaces in 
such a way as to develop the maximum power of the boiler at all times. 
This fact applies to refining furnace boilers in particular because of the 
intermittent nature of the refining operation. 

The gases from reverberatory furnaces are laden with dust and fume, 
which tend to deposit on the boiler tubes. Consequently the boiler 
and its setting must be arranged to permit frequent cleaning of the tube 
surfaces, and the removal of the dust which accumulates under the 
boiler. Fumes such as As 2 O 3 and Sbo0 3 tend to condense on the cool 
boiler tubes, and if the resulting deposit of dust and fume is not re- 
moved at frequent intervals it insulates the boiler and cuts down its 
efficiency. The tubes are generally cleaned by means of soot blowers 
supplemented by hand-operated lances of high-pressure air or steam. 
The dust which collects beneath the boiler is removed through clean-out 
doors. The first two entries in Table 1 give the analyses of waste-heat 
boiler dust at Anaconda; note the large amount of arsenic and antimony 
in these products, caused by the condensation of the volatile oxides on 
the cold boiler tubes. 

The waste-heat boilers installed on the reverberatory furnaces at the 
Douglas smelter 8 are located directly over the skimming end of the 
furnace; the boiler tubes are exposed directly to the molten bath in the 
furnace, and there is no damper between the furnace and boiler. 
Waste-heat recovery has been greater than was ever attained with the 
conventional system of boilers and furnace separated by flues largely 
because of the heat which reaches the boiler tubes by direct radiation 
from the incandescent bath of slag. Side skimming of slag is neces- 
sary in these furnaces because of the chilling effect of the boiler directly 
over the slag bath at the end of the furnace, and also to avoid fouling 

*McDaniel, L. L., New Reverberatory Waste-Heater Boiler and Power Plant 
at Douglas Smelter: Am. Inst. Min. & Met. Tech. Paper 996 (Metals Technology), 
February 1939. 



WASTE-HEAT BOILERS 



233 



of the slag by the high-grade sintered dust dislodged during hand 
lancing to clean the boiler tubes. 

Figures 1, 2, and 3 illustrate the waste-heat boiler installations on the 
smelting furnaces at Anaconda and United Verde (Clarkdale) and 
on the refining furnace at Great Falls, Montana. A brief summary 
of the practice at each plant is given below. 




(Barnard and Tryon, Am Inst. M\n *fr Met Eng Trans , Vol 106, p 232, 19SS) 

FIG. 1. Arrangement of Waste-Heat Boilers on Reverberatory Furnace, Anaconda 

Anaconda. The use of boilers in tandem at Anaconda has been 
necessitated by limitations in building space; these tandem boilers 
are smaller than would be the case if single boilers were used. The 
boilers are located between the rcverberatories (Fig. 1) and a pair of 
boilers may take the gases from the furnace on either side. About 
80 per cent of the steam is generated in the first boiler in this tandem 
arrangement. The small boilers in tandem do not require the use of 
baffles to slow down the passage of the hot gases. These baffles are 
used in some cases on waste-heat boilers to permit more efficient ab- 



234 SMOKE AND GASES 

straction of the heat in the gases; their principal disadvantages are 
(1) the increased resistance to passage of the gases and resulting loss 
of draft, and (2) the increased accumulation of accretions on the 
boiler tubes. 

The accretions on the boiler tubes result from the mixing of fine 
dust with arsenic trioxide and other volatile compounds which con- 
dense on the relatively cool boiler surfaces, and the deposit formed 
is rather sticky and can be removed satisfactorily only by use of a 
compressed air lance. Lancing is necessary about six times during 
an 8-hour shift. 

TABLE 2 
WASTE-HEAT BOILER DATA AT ANACOND\ 

Amount of fuel (natural gas) per 24 hours 2,742,000 cu ft 

Calorific power of fuel (gross) 1021 Btu/ru ft 

Make of boiler Stirling 

Average boiler rating 600 hp 

Type of boiler installation Tandem 
Equivalent evaporation from and at 

212 F per 24 hours 749,267 Ib 

Average steam-gage pressure 124 .5 Ib sq in. 

Average feed-water temperature 72 7 F 

Horsepower developed 904 9 hp 

Horsepower, per cent of rating 150 8 
Equivalent evaporation from and at 

212 F per 1000 cu ft of gas burned 303 10 Ib 
Percentage of total heat absorbed by 

boiler (based on gross Btu) 2S 81% 

Average gas temperature at boiler mlot 2050 F 

Average gas temperature at boiler exit cS10 F 

Solid material smelted per furnace day 612 27 tons 

United Verde. Two M-26 Stirling typo boiler- servo each roverbera- 
tory furnace (Fig. 2), each is rated at 713 boiler horsepower. These 
boilers are set at the same elevation as the tapping floor of the furnace 
so that the wavte-heat gases enter at the top of the front bank of 
tubes, necessitating the inversion of the u^ual baffling arrangements 

The boilers are equipped with valve-in-head soot blowers which 
keep most of the dust blown off the surface of the tubes, but it is neces- 
sary to supplement the soot blowers by high-pressure air lances inserted 
by an attendant through side doors in the boiler setting Pulverized 
coal is used for firing, and the coal ash and dust which accumulate in 
the flues between the furnaces and the boiler must be cleaned out by 
hand an operation which requires the greater part of an 8-hour shift 
during each 24 hours. 



WASTE-HEAT BOILERS 




236 



SMOKE AND GASES 



TABLE 3 
WASTE-HEAT BOILER DATA AT UNITED VERDE 

Amount of coal burned per day 112.58 tons 

Calorific power of coal as purchased (gross) 10,790 Btu/lb 

Calorific power of coal as burned (gross) 11,520 Btu/lb 

Feed-water temperature 113 F 

Temperature of steam 528 F 

Steam-gage pressure 178 Ib/sq in 
Distribution of heat in coal as burned: 
To smelting, conduction, convection, and 

radiation losses, etc. 43 0% 

Transferred to steam 43 6% 

Leaving in stack gases 13.4% 

Great Falls. Figure 3 shows a waste-heat boiler installation on a 
refinery furnace at Great Falls, Montana. This plant has three re- 
fining furnaces and each is provided with a boiler; two of the boilers 




WS^f^^^^^^ 

i- ^JMW?////////////////////////////////////////////, 

f ^ 



D 



1 

(BardweU, Am Inst Aftn <fc Met Eng Trans , Vol 106, p 886, 1933) 

Fia. 3. Waste-Heat Boiler on Refining Furnace, Great Falls. 

are Stirling class A-30, rated at 400 horsepower, and the third is a 
class A-21 Stirling boiler at 300 horsepower. The boilers are not 
baffled. 

Each furnace has a 125-foot steel stack which takes the gases as 
they leave the boilers and a bypass flue so that the gases can be 
diverted around the boiler if necessary (Fig. 3). The boilers are 



RECOVERY OF DUST AND FUME 



237 



ordinarily operated at 75 pounds steam pressure with feed water at 
60 F; from 36 to 40 per cent of the calorific power of the fuel is 
absorbed by the boilers. 

Table 4 shows the rate of steam production at different stages of the 
refining cycle for three separate furnace tests. Note the considerable 
variation in steam production during the different stages. 

TABLE 4 

RATE OF STEAM PRODUCTION, IN BOILER HORSEPOWER 
AT GREAT FALLS REFINERY 





Oil Fuel, 


Gas Fuel, 


Gas Fuel, 




400-hp Boiler 


400-hp Boiler 


300-hp Boiler 


Charging 


295 


178 


146 


Melting 


395 


290 


306 


Skimming 




310 


283 


Rabbling 


a 2H8 


298 


134 


Poling 


368 


573 


630 


Casting 


260 


441 


282 


Average 


337 


322 


312 



Includes skimming 

Other Methods of Heat Recovery. Attempts have been made to 
abstract more heat from watte gases by the use of economizers or 
recuperators which preheat the cold air used for combustion; these 
devices have been used on reverberatory ga^es after passing through 
the boilers and also on roaster gases. Although thcbe economizers 
have been used successfully in some places, they have not been 
universally adopted as have the waste-heat boilers. 

RECOVERY OF DUST AND FUME 

The common methods employed at copper smelters for removing 
suspended material from smokes arc: 

1. Collection of dust beneath flues, expansion chambers, balloon flues, 
and boiler settings. 

2. Filtration through cloth bags in " bag houses." 

3. Electrostatic precipitation in Cottrell t renters. 

Flues, Expansion Chambers, Etc. The principle involved in the 
recovery of dust and fume is simply that as the smoke is cooled by 
expansion in large chambers and by radiation and convection losses, 
its velocity and carrying power diminish. The cooling permits con- 
densation of fume, and particles of fume and dust settle to the bottom 



238 



SMOKE AND GASES 



of the chamber by the action of gravity. Only the larger particles 
will settle out in this manner, and the finer particles of dust and 
fume remain in the smoke stream from which they must be removed 
by other methods. 

All boiler settings, flues, etc., through which dust-laden gases are 
passing are equipped with clean-out doors or hoppers to permit removal 
of the accumulated dust; and the dust is cleaned out at intervals 
which depend upon the rate at which it collects. 




(Courtesy Tray lor Engineering and Manufacturing Company) 

FIG. 4. Converter Hood. 

Bag Houses. A bag house is a filtering chamber containing a 
number of cotton or woolen bags made of specially woven cloth. The 
bags are about 18 inches in diameter and 30 feet long; they are sus- 
pended vertically by means of a thimble at the top of each bag. The 
lower ends are connected to the gas intake, and the dust-laden gases 
are forced to enter the bags at the bottom and escape through the 
meshes of the cloth. Dust and fume are caught and held inside the 
bag, and the cleaned gases pass through. 

Woolen bags have longer life than cotton bags but are more expen- 
sive. Hot gas cannot be cleaned in bag houses because the heat destroys 



BAG HOUSES 



239 



the fabric gases should not be hotter than about 270 F (132 C) 
as a maximum and in many cases the limit would be 200 to 215 F. 
Acid smokes cannot be treated in bag houses because the sulfuric acid 
attacks the fabric and soon destroys it. 

The deposit which collects in the bags is removed by cutting off 
the gas entering the bag and shaking the bag vigorously a reverse 
current of gas may be drawn through the bag while it is being shaken to 



Motor and Reducer 
for Bag Shaking 



Flue from Blast 
Furnaces 



Van- 60,000 Cu. Ft/ Mm. @ 3Jr" 

Dust Chamber 



2- Stacks 5'x 10'x35' High 



s Steel Grating 
at Shaker Floor 

64' 6' 

PLAN 




ELEVATION SHOWING FAN 
AND CONNECTIONS 

FIG. 5. 



Damper Air Cylinder 
CROSS SECTION OF BAG HOUSE 
Bag House Layout. 



aid in loosening the cake. The collected dust and fume drop into 
hoppers beneath the bags. In most bag houses the bags are enclosed 
in steel compartments each of which holds about 12 bags, and the 
bags are shaken by means of an automatic hammer which raps the 
steel members supporting the bags. 

Bag houses offer considerable resistance to the flow of gas, and the 
draft loss will usually be from 3 to 6 inches of water gage. Auxiliary 
" booster " fans are generally used to handle gases passing through 
bag houses. 



240 SMOKE AND GASES 

Basic smokes are most commonly filtered in bag houses, and they 
are widely used to recover lead and zinc oxide fumes. The smokes 
produced in copper smelting are generally quite acid, so the use of bag 
houses in copper smelters is not as prevalent as their use in lead and 
zinc metallurgical works. When it is desired to filter acid smoke in a 
bag house it is possible to neutralize the smoke by introducing pul- 
verized lime into the smoke stream (Sprague process). This does not 
neutralize all the acid in the smoke, but the lime does collect on the 
bags and neutralizes the free acid which would otherwise destroy the 
fabric. At Tooele, Utah, 9 where bag houses were used to catch the 
fumes produced in converting leady matte, there was originally enough 
ZnO in the fume to neutralize the acid, but when the converting process 
was altered so that the zinc no longer entered the fume, it became neces- 
sary to add dehydrated lime to the gases just before they entered the 
bag house. 

Cottrell Treaters. The Cottrell process for removing suspended par- 
ticles from smoke utilizes the fact that if an electrostatic charge can 
be placed on these particles, they can then be attracted to an electrode 
carrying the opposite charge. Commercial Cottrell treaters are large 
chambers containing positive and negative electrodes; the positive 
electrodes have a large surface area and small radius of curvature as 
compared with the negative electrodes, and the dust and fume are 
collected on the positive electrodes. The positive electrodes are usually 
pipes or plates; the negative electrodes are wires or chains. The 
positive electrodes are grounded and the potential difference between 
positive and negative electrodes will be from 25,000 to 65,000 volts. 

Electrostatic precipitation requires two definite electrical con- 
siderations (1) there must be a unidirectional flow of ionizing cur- 
rent from one electrode, and (2) a high-potential static field must be 
maintained between the two electrodes. The discharge electrode (wire 
or chain) has a small radius of curvature and a small surface area, 
and hence it is possible to maintain a high density of charge on its 
surface. The potential gradient at its surface is sufficient to disrupt 
the neutral electrical state of the neighboring gas molecules and convert 
them into charged or ionized molecules with a charge of the same sign 
as that of the discharge electrode. As soon as a molecule becomes 
charged, it is subjected to the electrical stress of the static field and it 
moves away from the discharge electrode and toward the collecting 
electrode (pipe or plate), carrying with it other uncharged gas mole- 
cules. This gives rise to the " electric wind " and amounts to a very 

9 Sackett, B. L , Converting Lead and Copper Matte at Tooele : Am. Inst. Min. 
& Met. Eng. Trans., Vol. 106, p. 132, 1933. 



COTTRELL THEATERS 



241 



low amperage electric current passing through the gas; the current is 
carried on gas molecules which are charged at one electrode and dis- 
charged at the other. 

If the gas between the electrodes contains suspended particles of 
dust or fume, these collide with the gaseous ions, become charged, and 
are attracted to the collecting electrode, where they are discharged; 
these particles form an adherent deposit on the collecting electrode. 

/ High Tension Line Carrying 
Rectified Current , 



Upper 
Header 



Lower 
Header 



Dust 
Laden 
Gases 
Enter 



Low 

Tension 

Line from 

Switch- 

board 





Hopper- 



Discharge Electrode 
Collecting Electrode 
Discharge Electrode 



/Collecting Electrode- 

Suspended 
Material Collects on 
Inside of Collecting 

Electrode and 
Drops into Hopper 

Weight 




Transformer 



Precipitator 



Rectifier 

(Courtesy Western Precipitation Company) 

FIG. 6. Diagram of a Pipe-Type Cottrell Treater Unit 



The discharge electrode may be either positive or negative, but prac- 
tically it has been found that precipitation is more effective if the 
discharge electrode is negatively charged; and this system is used in 
all commercial Cottrell treaters. 

Practically any type of suspended material can be removed from a 
gas stream by Cottrell treaters, and the method has wide applications. 
It will remove all dust and fume found in copper smelter smokes and 
is the only method that will satisfactorily remove free H 2 S0 4 (or S0 3 ) . 
There is no important copper smelter in the United States or Canada 
which does not employ Cottrell treaters. 

The accumulated deposit which adheres to the pipes or plates is dis- 



Plate Type-2 Units 3 Sections in Series 




10 Feet 

Submerged Pipe Type-3 Sections in Parallel 
(Welch, Am Inut Mm A Met Eng Trans., Vol. 106, p. 818, 19S3) 

FIG. 7. Sketch Illustrating Nomenclature in Gas-Treating Parts of an Electrical 

Precipitation Installation. 

242 



COTTRELL THEATERS 



243 



lodged by rapping the electrodes at intervals with automatic hammers; 
the deposit falls into hoppers at the bottom of the treater chamber, 
from which it is removed at intervals. 

Smokes are classed as conducting or non-conducting according to 
the nature of the deposit formed on the collecting electrode; con- 



Coll Springs, 
for Ground 
frame Supp 



Cast Iron Weights-10* Each 
JIL- 

Baffle Plate 




A 



\ 



FIG. 8 Section of a Plate-Type Cottrell Treater. 

ducting smokes may be treated directly by the Cottrell process, 
but non-conducting smokes must first be conditioned. The non- 
conducting smokes are usually basic smokes, and common examples 
are smokes containing fumes of lead or zinc oxides; tnese substances 
form an insulating blanket on the collecting electrode and prevent the 
rapid discharge of the positively charged particles. This means that a 



244 SMOKE AND GASES 

positive charge accumulates on the electrode and does not leak off to 
the ground; this diminishes the static potential between the electrodes 
and cuts down the efficiency of the operation. 

Conducting smokes contain suspended matter (S0 3; H 2 S0 4 , H 2 0, 
and others) which renders the deposit conducting and permits rapid 
discharge of the particles to the ground. Acid smokes are conducting 
because only a small amount of H 2 S04 is needed to make the deposit 
a conductor. Non-conducting smokes are made conducting by adding 
a suitable conducting medium, and the most common method is humid- 
ifying or treating the smoke with a water spray the water droplets 
collect in the deposit and make it a conductor. A non-conducting dust 
can usually be made satisfactorily conducting if the collected dust 
contains from 2.5 to 4.0 per cent moisture. 

Most copper smelter smokes are conducting because many of them 
are acid, and many contain considerable moisture from the roasting or 
smelting of wet concentrates. Cottrell treaters are generally more 
suitable for treating copper smelter smokes than are bag houses, al- 
though bag houses are used for certain special purposes. 

Table 5 gives the essential data on the Cottrell installations at 
several representative copper smelters. 10 In operation, the gas passes 
through the treater between the collecting electrodes, and the dust is 
collected on them. The efficiency of collection depends upon the 
rate at which gas is passed through the treater with adequate time, 
practically all of the suspended matter can be removed. Commercial 
units usually show better than 90 per cent recovery. Capacities and 
efficiencies for several copper smelter installations are given in Table 5. 

TREATMENT OF RECOVERED DUST AND FUME 

Treatment given to the dust and fume collected in copper smelters 
will depend upon the composition of the dust, but usually the dust is 
simply charged back into the smelting circuit. Most dust is fed into 
the reverberatories, but some is also charged into roasters or converters. 

The principal byproduct from smelter fume is " white arsenic," 
As 2 3 ; practically all the world's supply of arsenic is a byproduct of 
copper and lead smelting. Where arsenic is present in any quantity 
in the smelter feed it tends to accumulate in the flue system because 
the lower oxide, As 2 3 is relatively volatile and is driven off in both 
the roasters and reverberatories. 

Crude arsenic-bearing dusts are subjected to repeated distillations 

10 Welch, H. V., Recovery of Suspended Solids from Furnace Gases: Am. Inst. 
Min. & Met. Eng. Trans., Vol. 106, p. 296, 1933. 



COTTRELL THEATERS 



245 







a 


"o 




1 | II '5 ^ 


t'3 


^ ^ 


1 ill 1 s 


.1 -1 


S fe 


g *!- 






$ fl - a , 


W 




Q! 0^2^g^ 
sS>O ivO^oco'CS 

o^ o V s oi ^0 oo r^ i 






S 


SO -2 


| 


1 


||l 


a 
2 







^ 


1 


1 


. s? 


o *o 

(N CO 


S 






t L t >o 




s I 


P O '5 t^ c3 

5" y o 


*3 c 


^ SL 




S 






5 1 

23 

s w 

Sw 


ifl 


ill 







S S 


* S 


.S * ^ 2 -2 8, 1 ^ S E - 


1^15 ilSj|| 


>J *- 

_. , -3 O 

W w o 

^ E g.2 


Ipl! iJl-illL 


'S ^ '.1 ^ ^ . 3 *i -^ 


S2 5 "5 


^_j C . f" t3 <H)co - ^ """' C "O * 


3 "p I 42 |o^c^| 


H ? ii 

1 
< 

Q 


i|ll^ . Cijillil 

O'^>>'2ti "g^aj^^^^n 

!3la!S If 111 III 


-822 i-*|-s| 

s>- 8 1 1 S^Ig- 

2g!s- :|.-&js 
a|.-s.i -ill-si'^ 

|3 5'^. 1 g-ft'532 


1 


u 


^s 1 


Q| 


1 


Jg.5 


<y 


rt o5 


1 1 ' 


C -d 

3 C 


S | 




"S rt 


2 - 


1 8 3 


55 


o 


Q 




2 o 


c 








g ^ 

5 


||| Jg 


| g 8 8 | - | 


3 S 


2 ^ S Q *"* 


** 2 3 "il a> * 


^? 


III i|| 


fc 8 - * s ij > Q 

S 9 S 'SS' > ^Q 

I a 1 1 1 g S 




(2 i S" <S 






i 1*1 










Hi 


o ^ ft 1 


S "^ 




1 8^8- 


g i 




o 5 


55 < 



246 SMOKE AND GASES 

and condensations until a commercially pure white arsenic is produced, 
and the residue is then sent back to the reverberatory furnace. 

The ore deposits at Butte contain the copper-arsenic mineral enargite, 
and as a result, a large amount of arsenic is recovered at the Anaconda 
smelter; we shall present a brief description of this plant to illustrate 
the methods used. 

Arsenic Recovery at Anaconda. 11 The principal feed to the arsenic 
plant consists of dust from the main flue leading to the stack and 
Cottrell dust from the treaters at the base of the stack. The flue dust 
is mixed with coal or flotation concentrates which reduce As 2 5 and 
arsenates to As 2 3 , and the mixture is treated in 6-hearth McDougall 
multiple-hearth roasters fired with gas burners on the third and fifth 
hearths. The gases from these furnaces pass into condensers, and the 
residue is either shipped back to the reverberatories or stock-piled and 
sold for its lead and bismuth content. The fumes from these furnaces 
pass into three condensers in series; these condensers are McDougall 
furnaces from which the interior hearths have been removed and 
baffles hung down the center. The rakes on the sixth hearth operate 
to discharge the condensed arsenic (together with some dust) , and the 
spent gases pass on to the main flue system. 

Cottrell dust mixed with coal or concentrate yields a crude arsenic 
containing 93 to 95 per cent As 2 3 , which is pure enough to go to the 
refinery. The low-grade flue dust, however, yields a more impure 
product, and this is usually mixed with flue dust to give a composition 
of 55 to 60 per cent As 2 3 and re-treated in the roasters; this gives a 
crude of about 94 per cent As 2 3 . 

Crude arsenic is treated in small batch reverberatories. The high- 
grade crude (93 to 95 per cent As 2 3 ) is heated on the hearth and the 
firing is regulated to give a temperature of 950 F. Arsenic-laden 
gases pass into brick chambers or arsenic kitchens, where the arsenic 
trioxide condenses to form commercial " white arsenic." About 1 % to 2 
tons of crude arsenic is treated at a time in the refining furnace, and at 
the end of each day the residue remaining is worked to the end of the 
furnace, where it drops into cars to be taken to the copper rever- 
beratories. 

Other Byproducts. Arsenic trioxide is the only important byproduct 
removed from the fume and dust in copper smelters; small amounts of 
lead and bismuth may be separated in the arsenic plant (as noted 
above). Unless these dusts contain enough arsenic to warrant special 
treatment, however, they are usually returned to the smelting circuit, 

11 Bender, L. V., and Goe, H. H., Production of Arsenic Trioxide at Anaconda: 
Am. Inst. Min. <fe Met. Eng. Trans , Vol. 106, p. 324, 1933. 



REMOVAL OF SO 2 FROM GASES 247 

and the contained impurities find their way out of the plant either in 
the slag or in the crude copper. 

In the treatment of the slimes produced in electrolytic refining, fur- 
nace operations produce dusts and fumes which contain valuable by- 
products. We shall discuss these in the next chapter. 

REMOVAL OF SO 2 FROM GASES 

A great deal of work has been done on the subject of removal of S0 2 
from gases in the last several years, and many methods have been 
developed for the recovery of SO 2 from gases by fixing it in the form 
of a non-volatile sulfur compound. One of the principal reasons 
for this is the fact that S0 2 discharged into the atmosphere in sufficient 
quantity is harmful to surrounding vegetation; the sulfur-bearing com- 
pounds formed by its removal will have some commercial value, but 
usually this itself is not a sufficient reason to warrant an elaborate 
treatment plant to recover sulfur dioxide especially from lean gases 
carrying up to 2 5 per cent S0 2 . 

Sulfur dioxide is a gas which is non-condensable at ordinary tempera- 
tures and pressures (its normal boiling point is 10C) so that it 
passes freely through bag houses, Cottrell treaters, etc., and it has only 
a limited solubility in w r ater. Commercial methods for extracting 
S0 2 usually depend upon converting the SO 2 into a compound which 
can be dissolved in water in relatively high concentrations. A summary 
of the various methods which have been used for this purpose is given 
m a recent publication of the United States Bureau of Mines; 12 but 
we shall mention only one method here the production of sulfuric 
acid as that is the only treatment that has been applied to any 
extent to copper smelter smokes. 

Production of Sulfuric Acid. 13 Sulfuric acid may be prepared from 
smelter gases containing S0 2 by the standard methods of manufacture, 
in which the gas is oxidized to S0,' 5 and dissolved in water to form 
H 2 S0 4 . We shall not enter into a discussion of sulfuric acid manu- 
facture but shall simply point out the relation of this method to the 
general problem of removing S0 2 from smelter smoke. 

In the first place, the S0 2 content of gas should be about 5 per cent 
by volume, or more, for acid manufacture. This means that the gases 
from converters, and some roasters and blast furnaces can be used 
for this purpose; reverberatory smoke, however, is too low in S0 2 . 

12 Roberson, A. H., and Marks, G. W M Fixation of Sulfur fnrn Smelter Smoke: 
U. S. Bur. Mines Kept Inv. 3415, October 1938. 

13 Fairlie, A. M., Sulfuric Acid Manufacture: Remhold Publishing Corporation, 
New York, 1936. 



248 SMOKE AND GASES 

Secondly, sulfuric acid is a fairly cheap commodity, and it does not 
pay to manufacture it in quantity when it must be shipped a great 
distance. 

Because of these two reasons we find that copper smelters ordinarily 
maintain sulfuric acid plants only when the sulfuric acid can be used 
directly in the plant ; and only a part of the smelter smoke is used for 
acid manufacture. Thus Anaconda manufactures sulfuric acid from 
a portion of the roaster gases and utilizes the acid in making phosphate 
fertilizer. Andes Copper Mining Company maintains a smelter for 
sulfide concentrates and a leaching plant for oxidized ore; part of the 
roaster gases are used to make the sulfuric acid needed by the leaching 
plant. The plant of the Tennessee Copper Corporation illustrates 
a special case (Chapter IV) ; here most of the gas produced from 
roasters, converter, and blast furnace goes to a sulfuric acid plant, 
and the sulfuric acid formed is one of the most important products of 
the low-copper, heavy sulfide ore. 

DISPOSAL OF WASTE GASES 

In modern copper smelters the gases are satisfactorily cleaned of 
suspended matter, and part of the S0 2 may have been removed. In 
order to render the waste gases as harmless as possible they are then 
discharged high in the atmosphere so that they may be highly diluted 
before they can diffuse back to earth. 

The question of damage caused by smelter smoke is still highly 
controversial, and it is very difficult to make positive statements about 
it; the amount of research being clone on methods to remove S0 2 
indicates that the " smelter smoke " problem is still of considerable 
importance. In most smelters, all the sulfur-bearing waste gases are 
led into a common flue and discharged through a high stack; shorter 
stacks may be employed for the direct disposal of gases which are free 
of S0 2 and suspended matter, such as those from cathode furnaces 
(Fig. 3) . Some of the tallest stacks in the world are found at copper 
smelters. 

Unless part of the sulfur is fixed in the form of sulfuric acid or 
other compound, all the sulfur that enters the smelter in ore or con- 
centrates must pass out in the waste gases. The amount of sulfur in 
the final stack gases at copper smelters will usually be about 2 or 
3 per cent by volume, although this may vary considerably. 



DISPOSAL OF WASTE GASES 249 

TABLE 6 a 
HEIGHTS OF SOME COPPER AND COPPER-NICKEL SMELTER STACKS 

Plant Height of Stack 

(copper) (feet) 

Anaconda, Montana 588 

Chinnampu, Korea 600 

Clarkdale, Arizona 430 
Copper Cliff, Ontario (copper-nickel) 510 and 554 

Tacoma, Washington 571 

Hurley, New Mexico 500 

U S Bur Mines Rept Inv 3415, p 42, 1938. 

TABLE 7 a 

RELATION OF ODOR INTENSITY AND PHYSIOLOGICAL EFFECT 
TO CONCENTRATION OF SO 2 IN AIR 

Concentration of SO2 by volume 

Parts per Million Per Cent Physiological Effect 

3 to 5 0003 to 0005 Faintly detectable by smell or taste. 

8 to 12 0.0008 to 0012 Slight throat irritation and tendency to 

cough, 

20 002 Very distinct throat irritation, coughing, 

constriction of chest, and smarting of 
eyes 

50 005 More pronounced irritation of eyes, throat, 

and chest, but possible to breathe several 
minutes. 

150 015 Extremely disagreeable, but may be en- 

dured for seveuil minutes 

500 050 Causes sensation of suffocation even with 

first breath 

u U. S. Bur Mines Rept Inv 34 If), p 45 



CHAPTER VIII 

ELECTROLYTIC REFINING 

INTRODUCTION 

The practice of refining copper electrolytically started late in the 
nineteenth century and has since assumed great importance in the 
metallurgy of copper. Of a total of 1,644,505,129 pounds of copper 
produced in the United States in 1937, * 1,548,857,307 (94%) was 
electrolytic copper. In Chapter VI we have already had occasion to 
refer to electrolytic refining in connection with the casting of anodes 
and the refining of electrolytic cathodes. 

As a brief introduction to the subject of electrolytic refining we 
shall quote a few excerpts from a paper by Walker 2 dealing with the 
history of its development. 

At the beginning of 1893 there were 11 electrolytic refineries in the United 
States, most of them using the multiple system. * * * 

The total production of these 11 plants in 1892 was very small; I have 
not been able to discover reliable figures of production. There were about 
thirty electrolytic copper refineries in the world that year, according to 
Titus Ulke, and the entire production of these refineries was 64,000,000 
pounds annually, as much as could be produced in the United States at the 
present time, 1931, in 6% days This shows conclusively what enormous 
strides have been made in the industry. * * * 

In the early nineties the electrolytic copper refiners found it practically 
impossible to produce copper of a standard quality regularly. We had a 
lot to learn. Sometimes the cathodes were tough, crystalline and pure; at 
others the product was distinctly inferior. I have seen cathodes in the 
tank room of the Baltimore Electric Refining Company covered with wide 
streaks of a brittle black deposit of copper containing impurities. It used 
to be said that when the cathodes were removed from the tank we might 
expect to take them out in sheets or with a shovel, which, of course, was 
an exaggeration. Still, at best the product was far from uniform, and on 
this account electrolytic copper did not command so high a price as the 
product from the Lake Superior mines, a standard brand of copper of excel- 

1 Minerals Yearbook, U. S. Bur. Minos, p. 90, 1937. 

2 Walker, A. L., Learning How to Refine and Cast Copper ; Choice of Methods 
in Mining and Metallurgy: Am. Inst. Mm. & Met. Eng., New York, 1932 (Seeley 
W. Mudd Series). 

250 



THEORY OF ELECTROLYTIC REFINING 251 

lent quality. The difference in price then was from % to % cent per 
pound, in favor of the latter. Electrolytic copper producers had a serious 
problem to solve. * * * 

It required years to convince the purchasers of copper that electrolytic 
copper of a high standard could always be regularly and uniformly produced, 
but the consumers finally were willing to concede that this class of copper 
was fully equal to the product coming from the Lake Superior district and, 
in fact, superior for electrical purposes on account of its higher electrical 
conductivity. From 1914 electrolytic copper has been used as the basis 
for official price quotations, Lake copper being sold at the same price to 
large consumers. * * * 

THEORY^ ELECTROLYTIC REFINING 

The theory involved in the electrolytic refining of copper is quite 
simple, and the principal obstacles that had to be overcome in develop- 
ing the process were the many details of practical operation. Let us 
begin by describing the copper coulometer, an instrument which 
operates exactly the same as an electrolytic refining cell and comes 
about as close to " perfect " electrolysis as is humanly possible. With 
this as a background we shall be better able to appreciate the less 
perfect cell used in electrolytic refining. 

The Copper Coulometer. 3 The copper coulometer is an instrument 
for measuring small quantities of direct current electricity, and it is 
similar in construction and operation to the even more accurate 
silver coulometer, by means of which the value of the standard ampere 
is defined the international ampere is the current strength which 
deposits 1.11800 milligrams of silver or 0.3294 milligram of copper 
from suitable solutions of salts of these metals in 1 second. 

The coulometer is simply a glass beaker containing an electrolyte 
into which dip two electrodes of pure copper. These electrodes are 
connected to the electrical terminals so that the current passes through 
the cell, and copper dissolves from the anode and deposits on the 
cathode (negative terminal). After the current has passed for a 
definite time, the cathode is removed, washed, and weighed, and for 
every 0.0003294 gram increase in weight, 1 coulomb (ampere-second) 
of electricity has passed through the cell. 

This last statement depends on Faraday's laws, which state: 

1. The quantities of substances set free at the electrodes are directly 
proportional to the quantity of electricity wfa'jh passes through the 
solution. 

3 Creighton, H. J., and Koehler, W. A., Electrochemistry: 3d ed., Vol. 1, p. 22, 
John Wiley and Sons, Inc., New York, 1935. 



252 ELECTROLYTIC REFINING 

2. The same quantity of electricity sets free the same number of 
equivalents of substances at the electrodes. 

From the second law we derive an electrical unit, the faraday, which 
is the quantity of electricity required to liberate 1 gram-equivalent of 
a substance; 1 faraday = 96,500 coulombs. 

Faraday's lawrare among the few generalizations ija physical science 
which are exact and to which there are no exceptions. This fact follows 
naturally from our conception of ionization and the nature of the flow 
of current through a second-class conductor (an electrolyte). An 
electric current is a directed flow of electrons, and such a current passes 
through a first-class conductor (such as a metallic wire) by the move- 
ment of the cloud of free electrons in the metal. There is no such 
flow of electrons in an electrolyte, but the " current " is carried by the 
free-moving ions found in the solution. For example, in the copper 
coulometer, the negative electrode gives up its excess electrons to 
neighboring copper ions, which become neutral copper atoms; at the 
positive electrode (anode) electrons are removed from the surface atoms 
of copper, which then become ions (CU++) and enter the solution 
as CuS0 4 . With the same number of electrons entering at the cathode 
as passing out at the anode, the anodic and cathodic reactions must be 
chemically equivalent; which in this case means that exactly the same 
weight of copper dissolves from the anode as plates out on the cathode. 

In other examples there will be different reactions at both anode and 
cathode, but regardless of the nature of the reactions the principle of 
equivalent reactions will hold. Electrolytic action is best regarded as 
two equivalent chemical reactions taking place at the electrodes; at 
the cathode, electrons enter the solution, and some substance is 
reduced; at the anode, electrons leave the solution and some substance 
is oxidized (has its valence increased). In the coulometer, copper ions 
are reduced at the cathode and metallic copper is oxidized at the anode. 

The anodic and cathodic reactions are therefore 

Cu - 2 (e) - Or"- 
and 

Cu ++ + 2 (e) - Cu 

where (e) represents an electron. 

Equivalent amounts of copper are involved, so there is no change 
in the composition of the electrolyte. If instead of^a copper anode 
we had used an insoluble anode, such as a strip of platinum, the metal 
would not corrode and the oxidation reaction at the anode might be 
represented thus: 

S0 4 --2(e)->S0 4 



THE COPPER COULOMETER 253 

This may not represent the exact sequence of reactions, because there 
probably is no such thing as a neutral SO 4 group ; however, it does in- 
dicate the end products of the reaction. The cathodic reaction would 
remain the same as before. The use of an insoluble anode would lead 
to the following effects. 

1. For every equivalent of copper plated on the cathode, one 
equivalent of H 2 SO 4 would be formed in the electrolyte, and l / 2 equiva- 
lent of 2 gas would be liberated at the anode. Hence as the 
electrolysis proceeded, the solution would become depleted in copper 
and enriched in free sulfuric acid. 

2. The net chemical reaction in the cell would be 

CuSO 4 + H 2 O -> Cu + H 2 S0 4 + |O 2 

When we had a soluble anode, however, there was no net cell reaction 
and consequently no change in the composition of the -electrolyte nor 
evolution of gas (O 2 ). 

Soluble anodes are used in copper refining, and we shall be concerned 
only with them in this chapter; insoluble anodes are used in the 
electrolytic extraction of copper from leach solutions, and we shall 
consider them in more detail in the next chapter. 

Let us now return to the description of the coulometer. The electro- 
lyte should be made as follows: 

150 grams of crystallized copper sulfate 
50 grams of sulfuric acid (sp gr 1.84) 
50 cc ethyl alcohol 
1000 cc distilled water 

The current density should be between 0.002 and 0.02 ampere per 
square centimeter of surface. The alcohol is used to minimize oxida- 
,ion of copper by air at the surface of the liquid the alcohol con- 
centrates in the surface layer and is itself slowly oxidized to acetone 
and acetic acid. The electrolyte is agitated by bubbling a stream of 
DO 2 gas through it. 

When the instrument is set up as described, the current efficiency is 
practically 100 per cent; i.e., all of the current passing is utilized in 
depositing copper and none is consumed by other reactions. The 
cathode deposit adheres tightly, and there is no chemical dissolution 
of the deposited copper in the electrolyte. Thus the weight of copper 
deposited can be taken as an accurate measure of the current. Let 
us now express the electrolyte composition and current density in the 
units commonly used in practice so that we can later compare these 
with the conditions which obtain in practice. Of course, alcohol is 
in commercial electrolytes, and we shall figure this as an 



254 ELECTROLYTIC REFINING 

equal volume of water; neglecting changes in volume due to solution, 
we find that the electrolyte contains approximately 

Copper, 33 grams per liter 

Free sulfuric acid, 44 grams per liter 

The current density will be from 18.6 to 186.0 amperes per square foot 
of cathode surface. 

Finally let us consider a list of the characteristics in which com- 
mercial electrolytic cells differ from the simple coulometer. 

1. Arrangement of electrodes. 

2. Construction of tanks. 

3. Composition of electrolyte. 

4. Composition of the anodes. 

5. Disposition of the impurities in the anode. 

6. Nature of the cathode deposit. 

7. Circulation of electrolyte. 

8. Current density and cell voltage. 

9. Purification of the electrolyte. 

10. Electrical equipment and conductors. 

REFINING METHODS 

There are two principal methods of copper refining the multiple 
or parallel system and the series system. In the multiple system 
separate anodes and cathodes are used and the cathode deposit is 
built up on a starting sheet made of refined copper. The series system 
employs no starting sheets, and the electrodes of impure copper serve 
as both anode and cathode the copper dissolving from one side of 
each electrode and the purified copper depositing on the opposite side 
of the adjacent electrode. All anodes and cathodes in a given cell have 
a multiple or parallel electrical connection in the first system; in the 
second system the electrodes in any one cell are in series electrically. 

The multiple system is more widely used than the series system. A 
brief comparison of the advantages and disadvantages of the two 
systems are given by Walker, 4 as follows (the Hayden and Nichols 
systems are both series systems) : 

Advantages of Multiple System : 

1. Ability to treat copper of any quality, no matter how impure or 
how rich in precious metal. 9 

2. Less loss in precious metals in the cathodes produced about 0.35 
per cent of the gold and silver in the anodes of the multiple system; 

4 Walker, A. L., op. cit , p. 73. 



THE MULTIPLE SYSTEM 255 

1 per cent in the Hayden system and 2 per cent in the Nichols process, 
due to the very long cathodes. 

3. Ability to handle electrodes and scrap in larger units and with 
less cost for labor. 

4. Requires less care in maintaining the purity of the electrolyte, 
as it is possible to effect a much better circulation. 

5. Cost of casting and preparing anodes much less, especially when 
compared with the Hayden system, where the plates must be rolled. 

Advantages of Series System: 

1. For a given amount of power more copper can be deposited. 
Although the leakage of current around the electrodes in this system is 
large, the voltage between plates is much less. The production in the 
Hayden system is about 140 per cent and in the Nichols cast anode 
system about 170 per cent of that in the multiple system per unit 
of power. 

2. Less carry of metals in process; electrodes are much thinner. 

3. Less scrap produced; about 6.7 per cent in the Hayden system, 
10 to 15 per cent in the Nichols, and 14 to 18 per cent in the multiple. 

4. Less tank room space required for a given output. Tanks can be 
placed closer together, and there are many more electrodes in each. 

5. Much less copper is required for busbars and conductors. 

The significance of these remarks will be brought out as we discuss 
the two methods. 

THE MULTIPLE SYSTEM 

Anodes. The anodes used in multiple refining are flat rectangular 
slabs of impure copper with two cast lugs at the top of each anode to _ 
support it as it hangs in the tank. Typical analyses of anodes are 
given in Table 5. Anodes are from V/ 2 to 2 inches thick, 35 to 40 
inches long, 28 to 36 inches wide, and weigh 500 to 700 pounds apiece. 
The anodes are usually transported to the tank house by means of a 
rack on which they are spaced just as they will be in the electrolytic 
cell; thus an entire load of anodes can be picked up by a crane and 
placed in the tank in a single operation. 

Cathodes. Cathodes are built up on starting sheets which are thin 
sheets of electrodepositcd copper made in special electrolytic tanks 
known as stripper tanks. These operate with regular anodes, but the 
cathode deposit is formed on plates of rolled copper. A current density 
of about 17 amperes per square foot is used, and : t takes about 24 hours 
to deposit a Vi6-i n ch layer on each side of the blank. These starting 
sheets are then stripped off the blanks (which had been oiled to make 
the sheets strip off more readily) , straightened, and fitted with a sup- 



256 



ELECTROLYTIC REFINING 




(Courtesy Anaconda Copper Mining Company) 



FIG. 1. 



Stripping Starting Sheets from 
Blanks. 



porting bar of copper. The starting sheet is usually fitted with one 
or two loops of sheet copper through which the supporting cathode bar 

extends (Fig. 2) . 

New starting sheets are usually 
placed in a tank one at a time, in- 
serted between the anodes which 
are already in position. While the 
sheets are still thin they have a 
tendency to curl and buckle, so that 
for the first day or so they must be 
inspected at intervals and removed 
and straightened if necessary. Af- 
ter 6 to 15 days the cathodes will 
have grown to the required size, and 
then they are removed, washed, 
and sent to the cathode furnace 
or sheared into smaller pieces for 
the market. Finished cathodes are 
1 '" - - ' about l / 2 inch thick, and their 

lengths and widths usually correspond to the anode dimensions; they 
weigh from 130 to 300 pounds each. From two to four crops of cathodes 
are commonly made from each set of anodes. 

Electrical Connections. Figure 4 illustrates the earliest and simplest 
form of multiple connections. If we think of the current as flowing 
from + to (the conventional picture, but actually opposite to the flow 
of electrons), the current enters the first tank through the + busbar. 
This is a heavy copper conductor of rectangular cross-section (about 2 
by 4 inches) running along one side of the tank ; the anodes are hung 
in the tank so that one lug of each anode rests on this busbar and the 
lug on the opposite side rests on an insulator. The cathodes are 
hung so that the cathode bars rest on the bus on the opposite side, and 
they are insulated from the anode bus. Thus the current enters on the 
anode bus, splits, and passes through the anodes, through the electro- 
lyte, on to the cathodes, and out through the cathode busbar. The 
extension of the cathode bus becomes the anode bus in the next tank. 

Anodes and cathodes in each tank are electrically in parallel or 
multiple, as the incoming current is evenly divided among the elec- 
trodes, and, except for minor differences, the voltage drop across the 
entire tank is the same as the voltage drop between any pair of 
anodes and cathodes. The tanks are electrically in series with one 
another. 
The current entering a tank may amount to several thousand amperes, 



JATHODES 



257 




Mining (V 

i. *2. in Tnuk. 

Note the methnd of attaeliliig caiit<>de 




(Courtesy Anaconda Copper Mining Company] 

FIG. 3. Lifting Cathodes out of an Electrolytic Refining Tank. 



258 



ELECTROLYTIC REFINING 



and a large bus is required to carry this current without undue heating. 
In the system illustrated in Figure 4 two full-length busbars are required 
for each tank plus the extensions connecting the tanks, and this means 
that a large amount of copper is tied up as busbars. Several modifica- 
tions of the original system have been developed to cut down the 
amount of copper required for conductors. 

Only one end of a busbar carries the total current entering the tank; 
for example the plus busbar of the first tank in Figure 4 brings in the 
entire current, but as the current passes down the bus, a certain amount 
is drawn off by each anode; near the opposite end of the tank the bus 
is carrying only a small fraction of the original current. Thus it is 



rm 



rm 



(From Creighton and Koehler, Electrochemistry, John Wiley and Sons, Inc , New York) 

FIG. 4. Simple Form of Multiple Connection. 

possible to use a tapered bus (Fig. 7) which contains less copper than a 
bar of uniform cross-section but is just as effective. Note that in the 
system illustrated in Figure 4 the cathode bus on the first tank would 
be tapered in the opposite direction to that of the anode bus. 

The Whitehead and Walker systems were devised to avoid the prac- 
tice of collecting the entire current from each tank on a heavy bus and 
then splitting it up again in the next tank. Figure 5 illustrates the 
Walker system, in which the tanks are placed side by side and the 



L 



Copjw conductor 




Insulate* 



(b) Section A-A 

(From Creighton and Koehler, Electrochemistry, John Wiley and Sons, Inc , New York) 

FIG. 5. The Walker System of Multiple Connection. 

cathode bars of tank No. 1, and the anode lugs of tank No. 2 rest on a 
common conductor set between the tanks; originally this conductor 
was a flat bar, but later a bar of triangular cross-section was used to 
obtain better contact. This conductor has a small cross-section as 
compared with a busbar because any given section carries oaly a small 
part of the total current. 



CURRENT AND VOLTAGE 



259 



(From Creighton and Koehler, Electro- 
chemistry, John Wiley and Sons, Inc , 
New York) 

FIG. 6. The Whitehead Single 
Contact System. 



The Whitehead " single contact " system (Fig. 6) does away with 
the conductor required by the Walker system, and the cathode bar in 
No. 1 tank rests directly on the lug of an 
anode in tank No. 2. In the original 
Whitehead system the anodes are cast 
with a triangle on top of one lug, and the 
cathode bar rests on this triangle; an- 
other development by P. K. Aubel utilizes 
a groove cast in the top of one lug 
(" Baltimore groove ") in which fits a 
wedge-shaped cathode bar. 

These modifications do not change the 
essential nature of the electrical con- 
nections electrodes in parallel in each 
tank and tanks connected in series. 

Current and Voltage. The current density commonly used in copper 
refining is 18 to 20 amperes per square foot of cathode surface, and 
the voltage drop per cell about 0.2 to 0.4 volt. The anodic and cathodic 
reactions balance each other almost exactly, so there is no " decompo- 
sition " or " chemical " potential and the voltage is simply that required 
to overcome the ohmic resistance of the electrolyte and the resistance 
of the contacts. 

Let us make a few calculations to illustrate the amount of current 
passing through an electrolytic tank, the rate of growth of the cathode 
deposit, and the reason for connecting the tanks in series. We shall 
assume that the current density is 20 amperes per square foot, that 
the cathodes have a submerged area 37 by 30 inches, and that each 
tank contains 28 anodes and 29 cathodes. This is the usual arrange- 
ment one extra cathode is used, and each cathode receives a deposit 
on both sides except the two end cathodes, which receive a heavy 
deposit only on the inner sides. In calculating the total cathode sur- 
face per tank, it is necessary to count both sides of all the inner cathodes, 
and one side of each end cathode. The total cathode area in this 
instance would be: 



07 y on 

^-^ 
144 



X 28 X 2 = 432 square feet 



and the total current per tank 

432 X 20 = 8640 amperes 

Now if the voltage drop per tank is 0.3 volt, each tank will require 
8.64 X 0.3 = 2.59 kilowatts of power, but commercial generators do 



260 ELECTROLYTIC REFINING 

not develop currents of such heavy amperage at such low voltages, and 
therefore it is necessary to connect a number of tanks in series. If 

190 
the generator could produce 8640 amperes at 190 volts, then = 633 

O.o 

tanks would be connected in series and served by the generator. 
Electrolytic plants require direct current, of course, and usually employ 
motor-generator sets to convert alternating current into direct current 
for the tanks. 

At 20 amperes per square foot and 100 per cent current efficiency, 
the weight of copper deposited per day per square foot would be 

63.57 20 

^ X 96^00- X 86,400 = 570 grams 

and this would mean a thickness (on one side) of: 



But the deposit forms on both sides, so the thickness will increase at 
the rate of 0.054 inch per day, thus requiring about 9 days to give a 
thickness of % inch. 

Current Efficiency. Although there are no exceptions to Faraday's 
law in the electrolysis of aqueous solutions such as these electrolytes, 
there may be several reasons why the deposit of metal actually formed 
is less than that calculated theoretically from the current flowing. The 
principal cause in copper refining is the fact that the solution picks up 
some oxygen from the air, and copper is slightly soluble in sulfuric 
acid solutions containing oxygen. As a result some of the copper 
deposited on the cathode is chemically dissolved by the electrolyte. 

The current efficiency is defined as the ratio (expressed as per cent) 
of the weight of the actual deposit to the weight calculated from the 
current flowing (by Faraday's law). Current efficiency in multiple 
refining is usually above 90 per cent and may reach 95 or 96 per cent. 

Table 1 gives some of the current and power requirements; these 
are calculated on a 93 per cent current efficiency and for voltages as 
indicated. 

The current efficiency varies inversely with the current density, the 
most efficient refining being done with low current densities. However, 
low current densities mean more plant space for a given capacity, and 
therefore the current densities used in practice are those which give 
the best balance between plant capacity and current loss. We have 
noted that multiple refining operations aim to maintain an efficiency 



ANODE IMPURITIES 



261 



greater than 90 per cent; series refining efficiencies are about 20 per 
cent lower, as we shall see later. 

TABLE 1 

ELECTRICAL REQUIREMENTS FOR COPPER REFINING AT 
93 PER CENT CURRENT EFFICIENCY 





Cell Voltage 


15 


20 


25 


30 


35 


0.40 


Ampere-hours per pound 














of copper 


411 


411 


411 


411 


411 


411 


Kilowatt-hours per 














pound of copper 


062 


082 


103 


123 


0.144 


0.165 


Pounds of copper per 














ampere-day 


0585 


0585 


0585 


0585 


0.0585 


0.0585 


Pounds of copper per 














kilowatt-day 


387 


292 


233 


195 


167 


145 



Anode Impurities. Copper anodes usually contain about 0.5 per 
cent total impurities, and these may include a number of base elements 
as well as precious metals (Table 2). As the object of refining the 
copper electrolytically is to remove these impurities and produce 
pure cathode copper, it is important to know how these impurities 
behave as the electrolysis proceeds. As the anode corrodes, one of 
three things happens; the impurities may: 

1. Dissolve with the copper and remain in solution. 

2. Dissolve and be reprecipitated chemically by ions in the electro- 
lyte. 

3. Remain undissolved and drop to the bottom of the tank as 
solid particles. 

Thus the contained impurities will collect principally in one of two 
places, either (1) dissolved in the electrolyte, or (2) in the deposit of 
anode mud, which accumulates on the bottom of the tank. Some of 
the suspended solid particles may be occluded in the cathode deposit 
and contaminate it; also some of the dissolved impurities may be 
plated out with the copper. 

As a general thing, those metals which are below copper in the electro- 
motive series do not dissolve but pass directly into the anode mud; all 
the precious elements are included in this gioup. Selenium and 
tellurium are present in the anode as selenides and tellurides of silver 
and copper; sulfur is present as copper sulfide, and oxygen as copper 
oxide. These compounds are insoluble and go into the anode mud. 



262 



ELECTROLYTIC REFINING 



Lead is soluble, but it is immediately precipitated as the insoluble 
lead sulfate, PbSO 4 . Arsenic and antimony are soluble, arsenic as 
As 2 3 , and antimony as H 3 Sb0 3 . Nickel, iron, and bismuth also 
dissolve in the electrolyte. Small amounts of NaCl or HC1 added to 
the electrolyte precipitate the bulk of the antimony as the oxychloride, 
bismuth as the oxychloride, and any dissolved silver as the chloride. 
The precipitation of antimony (and bismuth, when present) is not 
complete, but the presence of chlorine ions in the electrolyte keeps the 
amount of dissolved antimony below the point at which it would pre- 
cipitate on the cathode. 

Table 2 gives the approximate distribution of the various impurities. 

TABLE 2 a 
APPROXIMATE DISTRIBUTION OF COPPER ANODE IMPURITIES 





In Anode 


In Elec- 




In Anode 


In Elec- 




Mud 


trolyte 




Mud 


trolyte 




(per cent) 


(per cent) 




(per cent) 


(per cent) 


Gold 


100 


None 


Arsenic 


40 


60 


Silver 


100 


None 


Antimony 


90 


10 


Lead 


100 


None 


Selenium 


100 


None 


Nickel 


5 


95 


Tellurium 


100 


None 


Iron 


None 


100 









a Creighton, H J , and Koehler, W A , op cit , p 163 

The Electrolyte. We have already given the approximate compo- 
sition in terms of its free acid and copper content. The electrolyte 
found in an operating cell, however, will contain small amounts of 
other substances chlorine added as a precipitant, and the soluble 
impurities from the anode. Table 3 gives the average composition 
range of these electrolytes. As a rule the total amount of dissolved 
impurities should be kept below 25 grams per liter. 

TABLE 3 a 
COMPOSITION RANGE OF COPPER REFINING ELECTROLYTES 



Free 

Copper 

Nickel 

Arsenic 

Antimony 

Iron 

Chlorine 

Specific gravity 



180 to 220 g/1 


38 to 


45 g/1 


6 to 


10 g/1 


4 to 


12 g/1 


0.4 to 


0.6 g/1 


0.2 to 


1.2 g/1 


0.020 to 


0.052 g/1 



1.240 to 1.280 



" Creighton, H. J., and Koehler, W. A., Electrochemistry, 3d ed., Vol. 1, p. 162, John Wiley and 
Sons, Inc., New York, 1935. 



PURIFICATION OF THE ELECTROLYTE 263 

The sulfuric acid is used to increase the conductivity of the electro- 
lyte, and the electrolyte is also heated to about 60 C as this increases 
its conductivity, makes a firmer and denser cathode deposit, and 
causes more uniform corrosion of the anodes. The heating of the 
electrolyte, however, increases the chemical corrosion of both anodes 
and cathodes. 

Circulation of the Electrolyte. It is necessary that the electrolyte 
be kept continuously in circulation to avoid the segregation of salts 
which would take place if electrolysis were to be conducted in a 
stationary electrolyte. Without circulation tho heavy copper sulfate 
would tend to become more concentrated near the bottom of the 
tank, the solution near the cathodes would be impoverished in copper 
ions, and the solution near the anodes would increase in copper content 
until the hydratcd copper sulfate would precipitate on the anode and 
impede the dissolution of the copper. 

The solution is drawn from the tanks and pumped by means of air 
lifts or centrifugal pumps to an elevated storage tank from which it 
feeds back into the electrolytic tanks by gravity. In some plants the 
tanks are set on different floor levels so that it is possible for the 
electrolyte to flow through several tanks by gravity. The group of cells 
served by a single pump or air lift is known as a circulation cascade. 
The flow through any one tank is usually from an inlet pipe at the 
bottom of the tank near one end to an overflow pipe or weir at the top 
of the tank on the opposite end. Somewhere in the circulation system 
will be the device used to heat the electrolyte; the heating is usually 
done by steam coils either in the sump tank or the overhead storage 
tank. Lead pipe is commonly used for transporting the electrolyte, 
but the connections which enter the tanks must be made of rubber 
or other insulating material to prevent current leaking out on the pipe 
line. 

Circulation commonly amounts to 3 to 5 gallons per minute flowing 
into each tank. This gives a gentle flow with little turbulence in the 
cells, so that there is not so much interference with the settling of the 
slimes. Bottom-to-top circulation as described above is mos* common, 
although this causes the electrolyte to flow opposite to the settling of 
the slimes; circulation systems with inlet at the top and outlet at the 
bottom have also been used. 

Purification of the Electrolyte. For best operation of the electrolysis 
it is essential that the per cent of soluble substances in the electrolyte 
be kept within certain limits; copper and free acid should be held at 
definite values, and the soluble impurities must be kept below certain 
critical percentages (Table 3). In the copper coulometer there is no 



264 ELECTROLYTIC REFINING 

appreciable change in the composition of the electrolyte, but in com- 
mercial refining tanks several factors affect the electrolyte composition. 

1. Evaporation from the exposed surface of the heated electrolyte 
removes water and hence increases the concentration of all dissolved 
substances. 

2. Soluble impurities which are not precipitated by ions in the 
electrolyte build up in concentration as the electrolysis proceeds. 
Nickel, arsenic, and iron are the principal elements in this group. 

3. The copper content of the solution may either increase or de- 
crease depending upon the nature of the anodes. With fairly pure 
anodes, the copper will tend to increase in concentration because of the 
chemical corrosion of both anode and cathode by the sulfuric acid 
plus dissolved oxygen; this will also decrease the free acid content of 
the electrolyte. If the anodes are relatively impure and contain 
larger amounts of nickel and iron, an appreciable portion of the 
current will be consumed in corroding or " ionizing " these impurities 
at the anode. Practically all of the current is plating copper at the 
cathode, which means that copper is being plated faster than it is 
dissolving, and the electrolyte becomes depleted in copper sulfate. In 
most refineries, however, the anodes are pure enough that this second 
effect is not noticed; as a general rule it will be found that dissolved 
copper tends to build up in the electrolyte. 

Regulation of the electrolyte composition is attained by " bleeding " 
a certain amount of electrolyte from the main circuit and replacing this 
by fresh " make-up " solution of the proper composition. The foul 
electrolyte is then treated to recover as much of the acid and metal 
content as possible. Several methods have been used for treating 
foul electrolyte, the choice depending upon the nature and amount of 
the impurities. These methods utilize (1) crystallization of the 
salts by evaporation of the liquid, (2) electrodeposition using insoluble 
anodes, or (3) some combination of (1) and (2). 

The purpose of crystallization is to remove the excess copper sulfate 
as crystals of CuS0 4 -5H 2 ("blue vitriol"). This is done by first 
neutralizing the free acid by agitating with air in the presence of 
metallic copper, evaporating part of the water in the liquid, and then 
allowing it to cool so that the crystals may separate. In electrodepo- 
sition with insoluble anodes, the copper content of the electrolyte is de- 
pleted as there is no copper dissolved from the anode. This is a 
relatively expensive method from the standpoint of power costs be- 
cause of the higher voltage and lower current efficiency it usually 
takes from 8 to 10 times as much power to separate a pound of copper 
using insoluble anodes as it does with soluble anodes. 



ANODE CORROSION 265 

In the description of refining plants which is to follow we shall give 
some examples of the methods used in purifying electrolytes. 

Electrolytic Tanks. The tanks used in electrolytic refineries are 
made of either wood or reinforced concrete; reinforced concrete is 
superior to wood and is used in most of the newer installations. All 
tanks are lined with sheets of antimonial lead containing about 6 per 
cent antimony ; this resists the corrosive action of the acid and copper 
sulfate in the electrolyte and protects the wood or concrete from attack. 
Series cells are from 8 to 14 feet long, 2% to 3% feet wide, and 
3% to 4 feet deep. The tanks must be sturdily constructed to support 
the weight of the electrodes. An average tank will contain about 
10 tons of electrodes and 3 tons of electrolyte (570 gallons) . 

Piers of concrete or brick of sufficient height to provide 8 or 10 feet 
of head room below the tanks are commonly used for support. Each 
tank is fitted with a drain in the bottom so that the anode mud can be 
sluiced out. The floors of tank house and basement are generally 
covered with asphalt or other acid-resisting material. 

Anodes and cathodes are hung in the tanks spaced closely together. 
The electrode distance is usually about 4 to 4% inches. This is the 
distance from center to center of two adjacent anodes, and it means 
that each anode surface is separated from the adjacent cathode by a 
space of about an inch. This space will remain relatively constant as 
the electrolysis proceeds, because as the cathode deposit builds up the 
anode thickness decreases at about the same rate. Too close a 
spacing of electrodes may result in current losses caused by short- 
circuiting due to bridging of the gap between anode and cathode by 
crystals growing out from the cathode; too wide spacing of electrodes 
means more power loss, because the resistance is greater through the 
longer column of electrolyte, and less efficient utilization of the tank 
space. The tanks are wide enough to allow about an inch between the 
electrodes and the side walls, and deep enough to give a space of 6 
to 8 inches from the bottom of the electrodes to the tank floor. In this 
space the anode mud collects. 

Anode Corrosion. The copper and impurities dissolve from the 
anode under the action of the electric current, and the portion below 
the surface of the electrolyte gradually becomes thinner. Just before 
the corroded anodes are removed from the tank there will be only a 
thin sheet of the original anode left. The anode scrap, therefore, con- 
sists of the lugs and upper part of the original anode plus a small 
amount of the submerged portion (Fig. 7). Impure anodes corrode 
unevenly, and as the anode becomes thin there is a tendency for 
pieces of it to slough off and fall to the bottom of the tank. To avoid 



266 



ELECTROLYTIC REFINING 



this the corroded impure anodes must be removed sooner than purer 
anodes. The weight of anode scrap will be from 6 to 15 per cent of 
the weight of the original anodes, depending on their purity. Anode 
scrap is washed free of adhering slime and returned to the anode furnace. 












I ;:' \ ; ':;*:; : : * r .r; Jl *" JJ| ||j 















FIG. 7. View in the Tank Room of an Electrolytic Refinery. 

The tanks are arranged in " nesta " of ten and are connected by the Walker system. Busbars are 

tapered. In the foreground is a car of anode scrap to be sent back to the anode furnaces. Anodes and 

cathodes are handled by the overhead cranes. 

The Cathode Deposit. Two things are of importance in deposits of 
cathode copper (1) the chemical composition of the copper, and 
(2) the physical condition of the deposit. 

The cathode carries a negative charge, and in the adjacent electro- 
lyte there are numbers of cations carrying positive charges copper 
and hydrogen ions especially, together with some nickel and iron. 
Theoretically, all the metallic atoms below copper in the electromotive 
series should not dissolve at all, and the atoms of metals above copper 
in the series should dissolve but should not be discharged at the 
cathode as long as copper ions are present. Practically, this rule holds 
fairly well; thus hydrogen ions are not neutralized to any extent 
at the cathode, as hydrogen is well above copper in the electromotive 



THE CATHODE DEPOSIT 267 

series. The theoretical rule could be expected to hold only if the 
voltage drop between anode and cathode were very small; this, how- 
ever, would mean an extremely small current (Ohm's law), and the 
deposition would not be rapid enough for commercial refining. 

With higher voltages there is a tendency for other metallic ions to be 
neutralized (nickel, iron, etc.) and form part of the cathode deposit, 
and the greater the concentration of these ions in the electrolyte, the 
more that will be discharged. The higher the current density (and 
consequently the voltage) and the higher the concentration of soluble 
impurities, the less pure will be the cathode deposit. The limits of 
current density and impurity content that we have mentioned before 
are those which experience has shown will yield a deposit of the purity 
required in electrolytic copper. 

In addition to the impurities deposited electrolytically, suspended 
particles of anode mud may become attached to the cathode and be 
mechanically occluded in the cathode deposit. The amount of im- 
purities thus carried into the cathode copper will depend upon the 
amount of the anode mud and the rapidity with which the particles 
settle out. Arsenic and antimony tend to form float slimes of basic 
compounds which are particularly troublesome because they do not 
settle rapidly. More impurities are occluded near the bottom of the 
cathode than at the top, because the suspended particles are more con- 
centrated at the bottom. 

When copper ions are discharged at the cathode, the neutral atoms 
assemble into crystals of copper, and if a crystal is already started, the 
atoms tend to line up in its crystal lattice rather than to form new 
crystals. The growth starts by the formation of a large number of 
small crystal nuclei closely spaced over the cathode surface, and the 
crystals grow in a direction normal to the plane of the cathode surface. 
As these crystals grow, the more favorably oriented crystals increase in 
size, and the other crystals cease to grow thus the crystals appear 
to " fan out " as they get away from the original surface and instead 
of a large number of small crystals we have a smaller number of 
large crystals. The thicker the deposit the coarser the crystals will 
be, and the more irregular the surface of the deposit. 

Figure 8 is a photomicrograph of a cross-section of a copper cathode 
showing the starting sheet. One boundary of the starting sheet is 
straight, and the other is somewhat irregular. The straight side was 
originally adjacent to the starting blank in the stripper cell. Note 
that the crystals are small on this side and that they grow larger as 
their length increases; the "outer" side of the starting sheet is ir- 
regular because of the coarser crystals. The deposit formed in the 




268 ELECTROLYTIC REFINING 

commercial cell builds up on the starting sheet and begins with the 
formation of a new set of crystals. These also grow in size and eventu- 
ally become much larger than those in the starting sheet because the 
deposit is much thicker. For this reason the surface of the finished 

cathode is much more irregular 
than the surface of the starting 
sheet. 

Because copper crystals grow 
normally to the depositing sur- 
face, a groove on the starting 
sheet will result in a line of 
weakness in the sheet of de- 
posited copper. This fact is 
often utilized by making a 
small groove in the surface of 
starting blanks about % inch 
from the outer edge, so that 
the deposited copper will part 

^ *, -^ j o ,- nn, * A i_ readily along the line of this 
Fia. 8. Magnified Section Through the J . B 

Center of a Copper Cathode. groove when the starting sheet 

is being stripped from the blank. 

Crystal growth is always greatest at sharp points and corners, because 
the current density is greatest here. This often results in the formation 
of " bumps " and " warts " on the cathodes, and these may grow large 
enough to short-circuit the anode and cathode. 

In the ordinary deposit the copper crystals are firmly interlocked, 
and the deposit is strong and tough. Low current densities, hot solu- 
tions, and the addition of a small amount of glue (or similar colloid) to 
the electrolyte all promote greater smoothness and toughness in the 
cathode deposit. Small amounts of glue are used in practically all 
electrolytes, and it has a pronounced effect on the physical proper- 
ties of the cathode deposit. The glue, however, increases the re- 
sistance of the electrolyte, and even the small amounts used may 
increase the resistance by 10 or 20 per cent. 

Brittle Cathodes.** The process for producing copper shapes by 
coalescence and extrusion of cathode copper requires that the cathodes 
be cut or broken into small pieces. The ordinary tough cathode 
deposit is unsuitable, and a process has been developed to produce 
a brittle cathode. 

6 Tyssowski, John, The Coalescence Process for Producing Semifabricated Oxy- 
gen-Free Copper: Am. Inst. Min. & Met. Eng. Tech. Paper 1217 (Metals Tech- 
nology), June 1940. , 



PLANT OPERATIONS 269 

The production of brittle cathodes resembles the standard multiple 
refining process in most respects tanks and electrical connections, 
use of salt to provide chlorine, rate of circulation of electrolyte, acid 
content of electrolyte, temperature of electrolyte, length of deposit 
cycle, size and spacing of anodes, and removal of the : node mud. In 
fact, the two processes differ only m that (1) a depositing blank is 
used instead of a starting sheet, the deposit is shakei; or knocked off, 
and the starting blank used again, (2) the composition of the electro- 
lyte differs in certain minute but important details. 

Glue is omitted from the electrolyte, and m its stead there is 
added a small amount of an embrittling agent. This is a reagent 
which shows a greater ability to wet copper than docs the electrolyte; 
it should form n film on the copper which has a definite insulating 
value. A mixture of corn oil, castor oil, gasoline, and carbon tetra- 
chloride has been found to be a suitable embrittling agent. 

The starting blanks of ^-inch polished cold-rolled copper are 
dipped in the embrittling mixture before being placed in the tanks. 
The current is then able to puncture the insulating film only at a 
number of widely spaced points widely spaced as compared to the 
number of crystal nuclei that would form on an un-oiled sheet and 
there are fewer points at which crystal growth can start. As the 
crystals grow, the embrittling agent in the solution coats their surfaces 
and prevents them from growing together and forming a coherent 
mass. 

After the cathode deposit has reached the proper thickness, the 
cathodes are removed, and the deposit removed from the blank by 
shaking or knocking it; the friable deposit is then broken to the 
desired size, and the blanks are used again for the next cycle. 

The method for producing brittle cathodes compares quite favorably 
with standard refining practice. The process requires somewhat more 
labor in applying the embrittling agent and removing the cathode 
deposit, and the rough surface of the cathodes increases the possibility 
of entrapping slime particles. On the other hand, because of better 
electrical connections at the cathode (use of solid connections instead 
of loops and rods), and the increased conductivity of the electrolyte 
caused by the omission of glue, the voltage required is lower. Brittle 
cathodes require only 75 to 85 per cent of the power required to pro- 
duce tough cathodes under similar operating conditions. 

Plant Operations. The important operations in an electrolytic re- 
fining plant are as follows: 

1. Charging of new anodes and starting sheets in the tanks. 
Anodes are handled by means of overhead cranes which are equipped 



270 ELECTROLYTIC REFINING 

to move an entire tankload at one time. Starting sheets are usually 
placed by hand (Fig. 2) . 

2. Removing cathodes when finished, and washing them with a 
water spray to remove electrolyte and adhering slime. All the 
cathodes in a tank are removed at one time by the crane. 

3. Removing the remnants of copper anodes, washing them and 
returning the scrap to the anode furnaces. At the time the anodes 
are removed the anode mud is flushed out of the tank and collected. 

4. Inspecting the operating tanks to discover and eliminate short 
circuits caused by growths on the cathodes or the warping of new 
starting sheets. 



THE SERIES SYSTEM 

Much of the previous discussion on the multiple system of refining 
applies equally to the series system, and in such things as electrolyte 
composition, composition of anodes, and nature of cathode deposits, 
there is no important difference between the two methods. Our dis- 
cussion, therefore will be concerned principally with the outstanding 
differences between the two systems. 

Intermediate or Bipolar Electrodes. If an anode and cathode are 
immersed in an electrolyte there will exist a potential gradient through 

the column of electrolyte be- 
"" tween the positive and nega- 
tive electrodes, and if another 
A + metal strip is inserted in the 

6 electrolyte between the two 

electrodes, a current will flow 

ft . T , ,. . ^ . , through it because the side 

FIG. 9. An Intermediate Electrode. b . . . ... 

nearest the cathode will assume 

a negative charge and the side nearest the anode will have a positive 
charge. The intermediate electrode has no metallic connection with 
either of the two end electrodes (Fig. 9). The charges on the various 
parts are, of course, relative, because there is a steady potential gradient 
in one direction. Thus the right side of the intermediate electrode A 
is positive with respect to the left side of A ; however, the right side of 
A is negative with respect to the anode, and the left side of A is positive 
with respect to the cathode. 

If our solution contained dissolved CuS0 4 and the electrodes were 
of metallic copper we would find that when the current was passed 
through it, copper would dissolve from the anode and plate out on the 
cathode. Also, copper would plate out on the right side of A, and 



Cathode 



CURRENT EFFICIENCY AND POWER CONSUMPTION 271 

copper would be dissolved from the left side. A large number of inter- 
mediate electrodes might be inserted in the tank, in which case copper 
would dissolve from one side of each and plate out on the opposite side 
of the adjacent electrode. The electrodes in a series refining tank 
are a number of intermediate electrodes, and only the two end electrodes 
are connected to the electrical circuit (Fig. 9) . 

Electrodes. There are no separate anodes and cathodes in series 
refining tanks, as each of the bipolar electrodes serves as both anode and 
cathode, and no starting sheets or special depositing sheets are needed. 
The impure copper is carefully cast into the required shapes, or the 
electrodes may be made by a rolling process which yields a smoother 
and more perfect electrode. As the electrolysis piocceds, the copper 
dissolves from one side of each electrode and deposits on the next 
electrode; the electrodes are removed when the transfer is practically 
complete and the small amount of unrefined copper adhering is stripped 
off and scrapped. 

Series electrodes are smaller than the anodes used in multiple 
system, weighing about 100 pounds each. The busbars or cables 
which bring the current into the cells are not nearly as heavy as the 
busbars used in the multiple system, as the amperage per tank is 
much less. 

Series electrodes as cast or rolled are commonly called anodes, al- 
though this nomenclature is not strictly correct. 

Tanks. Series tanks are usually somewhat larger than multiple 
refining tanks but are of the same general construction. They also are 
constructed of wood or reinforced concrete, and the concrete appears 
to be the more satisfactory material. An important difference is the 
tank lining. A metallic lining cannot be used because this would by- 
pass a part of the current around the electrodes, so a non-conducting 
lining is necessary. 

Current Efficiency and Power Consumption. The overall current 
efficiency in the series process is lower than is common in multiple re- 
fining, generally in the neighborhood of 70 to 80 per cent. In spite of 
the lower current efficiency the series process generally yields a greater 
weight of refined copper per unit of electric power. This is dut, to the 
lower voltage required ; from 10 to 30 per cent of the total voltage drop 
in a multiple cell is due to the contact losses at the points where the 
anodes and cathodes rest on the conductors which carry the current 
these contacts are not needed in series refining. 

In most other respects there is no great difference between the two 
systems, and in another section we shall present a description of a 
plant using the series system which will illustrate the important details. 



272 ELECTROLYTIC REFINING 

TREATMENT OF ANODE MUD OR SLIME 

The weight of anode mud or slime produced in electrolytic refining 
of copper will range from 0.5 to 3.0 per cent of the original anode 
weight, depending on the purity of the anode copper. In addition to 
the elements we have already mentioned, a certain amount of metallic 
copper enters the slime. Part of this is in the form of small nodules 
which break off the cathode deposit, and part of it comes from the pre- 
cipitation reaction: 

2Cu+ - Cu + Cu++ 

This is due to the solution of a small amount of copper as cuprous ions 
which exchange charges to form a cupric ion and a neutral copper atom. 
Table 4 gives the average composition range of copper anode slimes 
as they come from the tanks. The principal value of these slimes is in 
their precious metal content. 

TABLE 4 a 
COMPOSITION RANGE OF COPPER ANODE SLIMES 

Gold 0.0548 to 6855%; 16 to 200 oz/ton; $560 to $7000/ton 

Silver 10.28 to 15 07%; 3,000 to 14,900 oz/ton 

Copper 16 to 24% 

Nickel 0.05 to 5 25% 

Lead 1 to 16%, 

Antimony 2. '6 to 8.0% 

Tellurium 08 to 6 0% 

Selenium 15 to 9 0% 

Arsenic 27 to 3 9% 

Bismuth 0.26 to 46% 

Iron 17 to 0.27% 

* Creighton, H J , and Kochler, W. A., Electrochemistry, 3d ed ,'Vol 2, p 168, John Wiley and 
SODB, Inc , New York, 1935 

The treatment of slime varies in different refineries, depending upon 
the analysis of the material; in general, however, the slimes are sub- 
mitted to three basic operations. 

1. Roasting to convert copper to copper oxide, and leaching with 
sulfuric acid to remove the copper. 

2. Subjecting the residue to a series of oxidizing fusions. This ox- 
idizes the base metal impurities and leaves a dore bullion of silver, gold, 
and platinum metals. Base metals pass into either slags or furnace 
fumes. 

3. Parting the dore to recover fine gold, fine silver, and platinum 
metals. 



ACID PARTING 273 

Roasting and Leaching. The slimes are first screened to remove 
any large pieces of copper and then filtered to give a dense cake con- 
taining about 35 per cent moisture. This cake is then roasted at about 
300 C, and the copper present is oxidized to CuO. The roasted slimes 
are then leached with hot sulfuric acid (10 to 15 per cent acid) in lead- 
lined kettles, and the leached slimes are filtered and sent to the dore 
furnaces. The solution from the leaching is returned to the tank house 
after being passed over metallic copper to precipitate any selenium 
and tellurium that may have dissolved. 

Oxidizing Fusion. The leached slime is melted down in a small 
reverberatory (dore) furnace, and the impurities are oxidized by air and 
by oxidizing fluxes, such as niter. Soda ash and silica are also used as 
fluxes. 

The sequence of operations such as skimming and adding of fluxes 
depends upon the composition of the mud being treated; the process is 
a batch operation and must be adapted to the particular charge being 
treated. If much lead is present the first skimming will consist largely 
of lead oxide (litharge) formed as the charge melts down. Further 
oxidation by means of air and oxidizing fluxes oxidizes the remaining 
base impurities, which are removed as slags. The gases from the dore 
furnaces are cleaned by passing them through spray chambers and Cot- 
trell treaters. The slags are re-treated to recover the impurities or sold 
for their metal content to lead smelters or other plants. Cottrell dust 
and the residue from the washing chambers are re-treated. Both the 
slags and the flue dusts contain considerable gold and silver as well as 
lead, selenium, and tellurium; all these refinery byproducts are too val- 
uable to be discarded. 

The metal remaining in the dore furnace is principally silver and gold 
containing small amounts of base metal plus any platinum metals in 
the original slime. This is then parted to separate and purify the gold 
and silver and to recover the platinum metals. 

Acid Parting. The dore is boiled in cast iron kettles in strong sulfuric 
acid (66 Be, or 96 to 98 per cent) to dissolve the silver and platinum 
metals. Gold remains undissolved, and the silver sulfate liquor is 
siphoned into a tank of water and boiled. The silver is precipitated in 
the form of crystals on copper plates hung in the tank, the copper (a 
less noble metal) going into solution. The silver crystals are collected 
and melted into bars of fine silver. 

If the platinum metals are present in sufficient quantity they must 
be recovered by separating them from the silver possibly by cast- 
ing the silver into anodes and refining it electrolytically. 

The use of acid parting is not common in large copper refineries 



274 



ELECTROLYTIC REFINING 



most of them employ electrolytic parting and refining methods. These 
are the Thum and Moebius processes for silver, and the Wohlwill proc- 
ess for gold. 

Thum and Moebius Processes. These two processes are electrolytic 
refining methods in which an anode of crude silver is electrolyzed and a 
deposit of pure silver is plated on the cathode. The two methods differ 
principally in the details of the cell construction. 



Glass rods^-^r 
(sirppoTts 
for anodes) 




Silver anodes (not shown jT Austin cloth on \ 
inplanulewabooe)A i top of-duch j 



Carbon or stainless steel cathode-^ 

"""s ^;-.^AJb 



'>>>>>>>>>>>>>>>>>>>>> 




(From Creighton and Koehler, Electrochemistry, John Wiley and Sons, Inc , New York) 

FIG. 10. Thum Cell for Silver Refining. 

The Thum cell is a shallow tank about 52 by 24 inches and 9 inches 
deep made of acid-proof stoneware or concrete lined with mastic. A 
slab of carbon or graphite covering the bottom of the cell serves as the 
cathode; stainless steel may also be used for the cathode. A wooden 
or stoneware basket (Fig. 10) with a bottom of glass rods serves to 
hold the anodes which are laid horizontally on the bottom of the basket 
with a piece of muslin or duck beneath them to prevent the anode slimes 
from falling through the bottom of the basket. The anodes are small 
slabs of impure silver about 8 by 12 inches; the cathode deposit forms 
on the cathode plate as loose crystals, which are raked out at intervals. 

The Moebius cell resembles a small multiple refining tank (Fig. 11), 
and the anodes and cathodes are hung from suspension bars. The 
anodes, which are about 14 by 5% inches and are about % inch thick, 



THUM AND MOEBIUS PROCESSES 



275 



are hung in canvas bags that hold back the anode slime. The cathodes 
are plates of stainless steel or rolled silver; the cathode deposit forms as 
loosely adherent crystals which are knocked off by wooden scrapers 
and collected in a basket in the bottom of the cell. The cell shown 
measures 24 by 26 inches; its depth is 22 inches. These cells are con- 
structed of the same material as the Thum cells. 




(From Creighton and Koehler, Electrochemistry, John Wiley and 6'ons, Inc , New York) 

FIG. 11. Moebius Cell for Silver Refining. 

The voltage drop in the Moebius cell is about 2.7 volts, and in the 
Thum cell it is 3 to 3.5 volts. This difference is due to the slime 
settling on the muslin diaphragm, which increases the resistance of the 
cell. Moebius cells require small floor space, use less electric power, 
and consume less nitric acid than the Thum cells; however, they pro- 
duce anode scrap which amounts to about 15 per cent of the weight of 
the anode and must be recast before it can be used in the Moebius cell. 
Anodes are completely consumed in the Thum cells because new anodes 
can be placed in the basket on top of the old ones, and all the fragments 
are thus dissolved. In most other respects there is little difference be- 
tween the two systems, and the remarks which follow apply to both. 

The anodes for these refining processes are made by casting th, bullion 
from the dore furnaces ; they contain about 95 per cent silver with the 
remainder mainly copper and gold. The silver and copper dissolve, 
and the gold remains behind as a slime. The cathode deposit contains 
99.9 per cent (999 fine) silver, which is melted in graphite crucibles and 
cast into bars. 

The cells operate with a current density of about 50 amperes per 
square foot. The electrolyte is a practically neutral solution of silver 



276 ELECTROLYTIC REFINING 

and copper nitrates containing about 60 grams of silver per liter and 30 
to 40 grams of copper. Part of the electrolyte is removed each day 
and replaced by fresh electrolyte in order to keep the impurities below 
the prescribed limits and to build up the silver content, as the electrolyte 
gradually becomes depleted in silver. Foul electrolyte is passed over 
copper to cement out the silver and then over metallic iron to precipitate 
the copper. 

The anode slime collects on the diaphragms surrounding the anodes. 
It contains all the gold, platinum, and palladium. This is removed, 
washed, and treated with boiling sulfuric acid to remove any copper 
and silver; it is then washed again, dried, melted, and cast into anodes 
for refining by the Wohlwill process. 

The Wohlwill Process. This process consists in electrolyzing impure 
gold anodes in a hot acid solution of gold chloride. Anodes are made 
from such materials as the anode slime from Thum and Moebius cells, 
and they will contain from 94 to 98 per cent gold. Cathodes are de- 
posited on rolled strips of pure gold, and the resultant cathode gold will 
be from 999.5 to 999.9 fine. 

The cells are made of glazed porcelain and are small in size because 
of the value of the electrolyte. Anodes and cathodes are suspended in 
the solution. The electrolyte contains 7 to 8 per cent gold as AuCl 3 
and 10 to 16 per cent free HC1; it is maintained at a temperature of 
70 C. High current densities (110 to 120 amperes per square foot) 
are employed to obtain rapid deposition and reduce interest charges on 
the gold in process; about 13 to 1.5 volts are required to give this 
current. 

Part of the electrolyte is removed daily to control impurities, and 
gold chloride must be added to replace the gold lost by depletion of the 
solution. Platinum and palladium concentrate in the electrolyte; but 
some gold, silver chloride, lead sulfate, and the other metals of the 
platinum group are found in the anode slimes. These are recovered 
by several different methods. 

The Wohlwill process utilizes an alternating current superimposed on 
the direct current used for the actual plating when the silver content of 
the anodes is high; this serves to prevent the silver chloride slime from 
adhering to the anodes. 

All three of these processes are used by the United States Mint as 
well as by commercial refiners. There are some differences in operat- 
ing methods, because the commercial refiners must put the metal 
through the process as rapidly as possible to minimize interest charges 
on the highly valuable metal and electrolyte; this problem does not 
confront the Mint. 



ELECTROLYTIC REFINING PLANTS 277 

ELECTROLYTIC REFINING PLANTS 

To point up the discussion of the methods used in copper refining we 
shall present brief descriptions of several multiple refining plants and 
one series plant. A good deal of the material will be given merely as 
a summary, but parts of the descriptions that have not been considered 
previously will be given in more detail. The two Canadian plants 
described are the most recently constructed refineries on the North 
American Continent. 

Raritan. 6 ' 7 The Raritan plant of the Anaconda Copper Mining 
Company is located on tidewater at Perth Amboy, New Jersey. The 
plant receives crude copper from Africa, South America, Mexico, and 
the United States, and all but about 6 per cent of this comes by water. 
Byproducts include silver, gold, platinum, palladium, selenium, tellu- 
rium, copper sulfate, and nickel sulfate. This plant has a production 
capacity of 45,000,000 pounds of refined copper per month. 

Electrical Plant. Power is produced by steam generated in three 
high pressure 1200-horsepower Stirling boilers which normally operate 
at 180 to 200 per cent of capacity at 385-pound pressure and deliver 
steam at 660 F. Four older B. and W. boilers of 760 horsepower 
each are kept in reserve, and the steam from the main boiler plant 
is augmented by steam from the waste-heat boilers on the eight copper 
furnaces. Power is generated as 60-cycle alternating current at 2300 
volts in three turbine-driven alternators, one of 3125 kilovolt-amperes 
rating and the other two of 5000 kilovolt-amperes each. 

Alternating current is converted to direct current at the tank houses 
by means of motor-generator sets. Tank house No 1 has two 2400- 
kilowatt sets; each of these consists of a synchronous motor (2650- 
kilowatt, 2300 volt, 720 revolutions per minute) direct-connected to 
two 300-volt, 4000-ampere direct-current generators. Thus the power 
supply to Tank house No. 1 amounts to 16,000 amperes at 300 volts. 

Tank house No. 2 is served by two 1920-kilowatt motor generator 
sets each consisting of a synchronous motor (2150-kilowatt, 2300-volt, 
720 revolutions per minute) direct-connected to two 240-volt 4000- 
ampere direct-current generators. These two provide current for the 
commercial cells, which amounts to 16,000 amperes at 240 volts. Tank 
house No. 2 contains all the stripper cells for making starting sheets, 
and these are supplied with 7000 amperes at 100 volts from a 1076- 
kilowatt rotary converter equipped for automatic constant-current 
control. 

6 Burns, W. T,, Refining Anaconda Copper at Raritan and Great Falls: Eng. 
and Min Jour , Vol. 128, No 8, p 306, 1929 

7 Raritan Copper Works, Pamphlet issued by Raritan plant. 



278 



ELECTROLYTIC REFINING 



_. E 

IE 

!" 
If 
$t 




o 

i 

& 



& 

G 

e 



1 



RARITAN 279 

Current density used is 18 amperes per square foot of cathode in the 
commercial cells and 17 amperes in the stripper cells. Voltage drop 
per cell is about 0.25 volt, and the current efficiency (as calculated 
from the current delivered to the tank house and the weight of cathode 
copper produced) is 92 to 95 per cent. 

Tanks. Tank house No 1 contains 1800 cells or tanks, and tank 
house No. 2 contains 1656 cells. There are 276 stripper cells for start- 
ing sheets, all in tank house No. 2, so that the total number of cells is 
3180 commercial cells and 276 stripper cells. 

Tanks are 9 feet 11 inches long, 2 feet 10 inches wide, and 4 feet 
deep. Most of the tanks are built of wood and are lined with %-inch 
6 per cent antimonial lead sheeting. The newer tanks are of lead- 
lined reinforced concrete ; all replacements are to be concrete tanks. 

Tank house No. 1 is divided into two electrical circuits of 900 tanks 
each. The commercial cells in tank house No. 2 are divided into two 
circuits, one of 660 cells and the other of 720 cells; stripper cells are 
on a separate circuit. The Walker system of electrical connections is 
used. 

Each tank house is served by eight electrically operated cranes for 
handling anodes and cathodes. Commercial tanks each contain 28 
anodes and 29 cathodes ; stripper cells contain 25 anodes and 24 starting 
blanks. 

Anodes. All anodes are cast into straight-line casting machines at 
the plant from four anode furnaces. 

Composition : 



Cu, 99.25% 
O 2 , 0.100% 
Ag, 30 oz/ton 
Au, 40 oz/ton 
As, 060% 
Sb, 022% 
Ni, 0.050% 


Pb, 0.053% 
Fe, 058% 
Bi, 003% 
S, 0.004% 
Se, 048% 
Te, 0.038% 



Weight: 525 pounds 

Size: 36 inches long, 28 inches wide, l| inches thick. 

Life: 30 days produce three crops of cathodes. 

Anode scrap: 10 to 12 per cent. 

Mode of suspension : cast lugs. 



Cathodes. The starting sheets are deposited on sheets of rolled 
copper in the stripper cells; the blanks are greased to prevent the de- 
posit sticking. At the end of 24 hours the blanks are pulled and an 
8-pound starting sheet is stripped from each side. To each starting 



280 ELECTROLYTIC REFINING 

sheet are riveted two small copper loops through which the cathode 
supporting bar is passed. 

Commercial cathodes weigh 160 pounds and are pulled at the end of 
10 days. These are washed and then melted and cast in one of four 
cathode furnaces. 

Composition of cast electrolytic copper (wirebars, etc. ) : 



Cu, 99. 94 to 99. 97% 

Ag, 0.0010% 

Au, 0.00001% 

S, 0.0020% 



2 , 0.02 to 0.05% 

Fe, 0.0025% 

Ni, 0.0015% 

As, 0.0015% 



Sb, 0015% 

Electrolyte. The electrolyte is circulated by means of vertical cen- 
trifugal pumps and passes through the tanks at the rate of 4 gallons per 
minute. 

Composition, in grams per liter: 

Cu, 45 Ni, 8.5 

As, 12 5 H 2 S0 4 , 200 

Sb, 0.4 Cl, 0.03 
Fe, 1.20 

Temperature: 55 C 
Specific gravity : 1 . 26-1 . 28 

Anode Mud. The anode mud produced will average about 0.6 per 
cent of the weight of anodes. The mud produced from Anaconda 
anodes will be as follows: 



Ag, 43 23% (12,610 oz/ton) 
Au, 234% (68 4 oz/ton) 

Cu, 13 86% 
Se, 1.46% 
Bi, 26% 



Te, 6 14% 

Fe, 22% 

Sb, 2 46% 

Pb, 3 96 

Ni, 0.27% 



As, 3.88% 

The Rantan silver refinery not only handles the anode mud pro- 
duced in its own plant but also anode mud from the Great Falls re- 
finery and dore bars from the lead refinery of the International Lead 
Company. The capacity of the silver refinery is 2,500,000 troy ounces 
of silver and 25,000 troy ounces of <gold per month. 

Slime is flushed out of each tank by removing a lead plug in the 
bottom, and the slime is carried through launders to a collecting tank 
in the basement, from which it is pumped to the silver refinery. This 
mixture of slime and electrolyte is thickened in settling tanks and the 
clarified electrolyte returned to the tank house. The slime is then fil- 
tered on an Oliver filter and given a light roast in an oil-fired furnace 



RARITAN 



281 



to break up selenides and tellurides and oxidize the copper. The 
roasted slime is then agitated in lead-lined iron kettles with hot 10 per 

RAW SLIME 
WITH SOLUTION FROM ELECTROLYTIC COPPER REFINERY 

I 





SETTLING TANKS 




' ' SLIME SETTLED OUT, SOLUTION OVERFLOWING 


1 1 

SETTLED SOLUTION RETURNED 
SLIME To COPPER REFINERY 

1 




FILTER FEE / TANK 




1 




*4 FILTRATE 


OLIVER FILTER 




FILTERED AND WASHED WITH WATER 




\ 




r 

RAW SLIME 

FOR THE 

PRECIPITATION 
OF SILVER 
FROM SOLUTION 

1 


1 


ROASTING FURNACE 


OXIDATION AT 700 F 


t 

ROASTED SLIME 

1 


* LEACHING TANKS 




Cu So 4 LEACH 
i< LIQUOR AND -* 
WASH WATER 




BOILED AND AGITATED 
WITH 15* H 2 So 4 SOLUTION 
AND WASHED WITH WATER 


* 

LEACHED SLIME 

I 




SHRIVER DIAPHRAGM PUMP 




PRECIPITATING 
TANKS 






FILTER FEED TANK 


SOLUTION BOILED 
IN CONTACT WITH 
COPPER SHEETS -t FILTRAT 


i 


E < MOORE OLIVER FILTER 


I I 
14-* Cu So 4 -<-J * > PRECIPITATE 
Qni MTiriM 


FILTERED AND WASHED 
WITH WATER 



TREATED SLIME 
TO PORE FURNACES 
(Masker, Am, InaL Mm. and Met Eng Trans., Vol 106, p 428, 1933) 

FIG. 13. Flowsheet of Treatment of Anode Mud or Slime. 

cent sulfuric acid to dissolve the copper. After again washing and 
filtering on a Moore filter the decopperized slime goes to the dore 
furnaces. 
The dore furnaces are small oil-fired reverberatories which hold 



282 



ELECTROLYTIC REFINING 



about 15,000 pounds of wet slime ; treatment of this charge takes about 
40 hours and results in the production of 3500 pounds of dore bullion 
containing 98.5 per cent silver, 1.0 per cent gold, and 0.5 per cent other 
metals. This dore is cast into anodes for the Thum and Moebius cells. 
A flowsheet of the Raritan plant shown in Figure 12 indicates the 
number of products removed in the dore furnace. Lead and antimony 
are collected in slags and selenium and tellurium are recovered both 

TREATED+ SLIME 



REVERBERATORY REFINING (DORE) FURNACES 

AND REFINING WITH HuCOl NiNOl'ANO AIR TO OORC 



ALKALINI 
SLAG 





SELENIUM-TELLURIUM 
PLANT 


1 JAW CRUSHER 1 Ult A V 1 !2 o " >/ \ JAW CRUSH 
ICRUSHED TO V^INCH] , r t *t** Lrrtc [CRUSHED TO i 


LEACHING VATS 1 "ART.N PLANT) tEAD^N^N^, 
LEACHED WITH WATER) LEAD REFINER 


DISSOLVING TANK 

""* CRYSTALS DISSOLVED 
IN WATER 


ALKALINE LEACHED 6 LA 3 




SCLENIOua ACID 










SELENIUM PLANT 
NEUTRALIZING TANKS 

BOILED AND NEUTRALIZED 
WITH CRUDE H, SO, OR ADDED 
ALKALINE LEACH LIQUOR 


COTTRELL EL 
PRECIPI7 

CLARIFICATION 
VOLTAGE OIREC 


TELLURIUM 
DIOXIDE 
PRECIPITATE 

* 


NEUTRAL SELENIUM 
SOLUTION 


* PR 

CLARIFIED ou 
OASES TO SL 
ATMOSPHERE Ctc , 


GASSING TANKS, mmftir 
ACIDIFIED WITH 10* -^s^ENf 
HCI AND OASSKO *, L "7'^ 
UMTM SO, TO WASTE 


[FILTER PRESS 1 
WASHED WITH WATERJ 


SETTLING 1 
SLUDGE SETTL 
WATER OVERF 


TO WASTC T RED SELENIUM 


CRUCIBLE HsTEAM DRYE 
FURNACE f 


3 wSam" U A T 8 H w w AVT c E R 


WATER PRE 
TO WASTE 


' MELTED WITH ilMECHANICA 
POWDERED COAL, II SCREEN 


" BLACK SELENIUM 




*1 ( 'iSTEAM DRVER ^ 


eVo 1 POWDER 
REWORKED | TO MARKE 


r l ""-ILL 1 


TELLURIUM ' 1 ' 
CAKES BLACK POWDERED SELENIUM 
TO MARKET TO MARKET 


ELE 

OX 
CRY 



FCUE 
OASIS 



MAIN FLUE 
CHAMBERS 

COOLED AND 
SETTLED 



SCRUBBER TOWERS 
OASES SCRUBBED WITH RECIRCULATCD 

WATER, OUST IN. GASES FORMING 
SLUDGE WHICH WITH EXCESS WATER 
OVERFLOWS FROM WATER SEAL TANK 



MOISTENED 



SLUDQf 



T OVERFLOW 

ICRUBBCR WATER 




WATER SLUDQE 

rASTE PRESS.CAKE 



COOLINO FLUE 



ROASTED 

SLUOOE 

SHIPPED TO 



(Mother , Am. Inst. Mm and Met Eng Trans , Vol 10G, p 4318, 1933) 

Fia. 14. Flowsheet of Furnace Refining and Recovery of Selenium and Tellurium. 

from the flue dusts and in the alkaline slags produced by the addi- 
tion of soda ash and niter. The amount of tellurium and selenium 
produced depends on the market for these elements. Although se- 
lenium and tellurium are present only as small percentages even in the 
anode slimes, the byproduction of these elements from copper refineries 
accounts for practically all the world's production; usually the supply 
has been greater than the demand. Most of the byproducts must be 
thrown into a commercially salable form or discarded; it is not possible 
in a refinery to keep impurities circulating in the system. 

Both Thum and Moebius cells are used for parting the dor6, and the 
cathode crystals are washed, melted in large graphite crucibles, and 
cast into standard thousand-ounce bars with a fineness of 999 + . The 
gold slime is leached with boiling sulfuric acid to remove silver, 
washed, and cast into anodes for the Wohlwill cells. The silver sulfate 



RARITAN 



283 



DOPE (SILVER-GOLD) ANODES 








* 




4 


ELECTROLYTIC CELLS-THUM & MOEBIUS 


DISSOLVING TANK 


STORAGE > OTHER 
TANK 


IN SILVER- COPPER NITRATE SOLUTION <"| 


NITRIC AGIO 

SILVER n j tr . 
N.TRATC ou> 
SOLUTION Muo 


CRYSTAL SILVER 




QOLD MUD ANODE 1 
4, SCRAP | 


WASH CARS 


WA 




* i 


WASHED WITH WATER 


IHED WITH WATER | 




WASH WASHED 1 
WATER CRYSTAL 1 
i SILVER T 


\ 
WASHED MUD 
I 


*~ 1 


WASH 
WATEI 


BOILING TANK 


CEMENTING TANK 


F 


ADJUST > 




BOILED WITH WATER J , 
AND DILUTE H 2 S04 vt 


BOILED IN CONTACT 
riTH COPPER SHEETS 




FINE SILVER BARS 




T TREATMENT X 
TREATED L|QUOR ' 
MUD J. SILV 
1 ^ > Got 
| * CEMI 


* 

ER. CuSCU 
o SOLUTION 
NT, TO COPPER 
JRE REFINERY 

ACK. 


To MARKET 








Li 






DEWATERED MUD 
4, FILTRATE 




OIL-FIRED KETTLE ^ 


BOILED WITH 60 BE H 2 SO4 "" 


GOLD SAND 


KETTLE LIQUOR 




FILTROS FILTER 




WASHED SA 


rJ 


r 

KATE 


CRUCIBLE FURNACE ^ 


T 

GOLD ANODES 








GOLD CATHODES 


P 

SAT 
ELEC 


| ,JL * 

e st?:r 

U..T.O I Sc , A 


} 

E 
P 


WASH TANK 


BOILED WITH WATER 


WA *WTER WASHED 


DISSOLVING TANK 1 


1 * CATHODES 




WASH TANK 


CRUCIBLE FURNACE 


EVAPORATED OFF WAS 


HED WITH WATER 


1 * 


1 t 1 

GOLD SILVER 
CHLORIDE CHLORIDE WASH 
SOLUTION RESIDUE SCRA 


4 

WASH 
EO WATER 
P 1 


^ to MARKET 




MECHANICAL ROLLS 


PLA 
PAL 
REC 


TINUM- FURNACE 

LAOIUM 1 


CATMODC SHEETS 


,OVERY 




CATHODE SHEETS 


\ 

PLATINUM 


I 

PALLADIUM. 
SPONQE 







(Mother, Am Inst Mm and Met Eng Trans , Vol. 106, p. 434, 19SS) 

FIG. 15. Flowsheet of Electrolytic Parting and Refining of Precious Metals. 

solution is cemented to remove the silver, which returns to the dore 
furnace. 

The Wohlwill cells produce gold about 999.75 *ine; this is melted in 
graphite crucibles and cast into standard bars. Platinum and palla- 
dium are recovered by working up the foul electrolyte from the Wohlwill 
cells. 



284 ELECTROLYTIC REFINING 

A complete description of the silver refinery at Raritan has been 
given by Mosher. 8 We shall not go into further details on this plant 
beyond presenting three flowsheets (Figs. 13, 14, and 15) , pages 281-283, 
which show these operations in more detail than Figure 12. 

Purification of the Electrolyte. Impure electrolyte from the tank 
house is treated in insoluble-anode liberator cells, to produce a copper- 
arsenic deposit and sulfuric acid. The copper-arsenic residue goes to 
the smelter, and the acid joins the electrolyte in the secondary metal 
tank house, which is an electrolytic refinery for the treatment of sec- 
ondary (scrap) copper. 

Great Falls. The Great Falls refinery 9 - 10 of the Anaconda Com- 
pany is located at Great Falls, Montana, about 200 miles from the 
smelter at Anaconda. It is smaller than the Raritan refinery and does 
not include a plant for treatment of slimes anode slimes are shipped 
to the Raritan refinery. The plant has a capacity of about 27,000,000 
pounds of refined copper per month approximately half that of the 
Raritan plant. 

Electrical Plant. The refinery is located near the hydroelectric 
plant of the Montana Power Company on the Missouri River, and 
power is purchased as alternating current at 6600 volts. The refinery 
substation contains seven synchronous motor-generator sets, each con- 
sisting of one 1730-horsepower 6600- volt alternating-current motor 
driving two 600-kilowatt 200-volt 3000-ampere direct-current genera- 
tors which furnish power for the electrolytic circuit. This gives a total 
of 42,000 amperes at 200 volts available for the tank house. Current 
density is high about 28 amperes per square foot of cathode and 
the voltage drop per tank is 0.4 volt. 

Tanks. The tanks are made of wood and are lead lined; inside di- 
mensions are 10 feet 3 inches long, 2 feet 10 inches wide, and 3 feet 9 
inches deep. Some of the tanks are of lead-lined reinforced concrete, 
and all replacements are to be concrete tanks. The plant is arranged 
in the two-cell system, i.e., a group of cells is arranged in double rows 
of ten, five in a cascade, with aisles between each double row. This is 
different from most other refineries which use the Walker or White- 
head systems requiring the tanks to be built in " nests " (Fig. 7) . The 
Great Falls arrangement requires more busbars and more floor space; 
however, each tank is an independent unit, and when a tank is in need 
of repairs a new or rebuilt tank can be put in place in a few moments. 

8 Mosher, M. A., Recovery of Precious and Secondary Metals from Electro- 
lytic Copper Refining: Am. Inst. Min. & Met. Eng. Trans , Vol. 106, p. 427, 1933. 

9 Bums, W. T., op. cit., p. 306 

10 Bardwell, E. S., and Lapee, R. J., Notes on Purification of Electrolytes in 
Copper Refining: Am. Inst. Min. & Met. Eng. Trans., Vol. 106, p. 417, 1933. 



GREAT FALLS 



285 



Worn tanks are continually being replaced by new ones, so that no 
wholesale replacing of old tanks need be made at any one time. 

The tank house contains 1440 commercial tanks and 90 stripper cells 
for making starting sheets. These are divided up as follows: 

1. There are four crane bays of 360 cells each served by seven 10-ton 
cranes of 60-foot span; a crane transfer at one end of the building 
makes it possible to transfer cranes from one bay to another. The 90 
stripper cells are not under the cranes. 

2. The tank house is divided into 12 sections of 120 cells each and 
the stripper section of 90 cells. Each section has a separate electrolyte, 
which is pumped by Pohle air lifts. 

3. The tanks (commercial and stripper cells) are divided into 
three electrical circuits of 540, 510, and 480 tanks. 

Commercial cells each hold 25 anodes and 26 cathodes. 

Anodes. Anodes are of the coped lug type and weigh 630 pounds 
each. Anodes are cast at the Anaconda smelter and shipped to Great 
Falls. The only anodes cast at Great Falls are those made from the 
anode scrap. 

Composition: Table 5 gives typical analyses of the anode copper 
treated at Great Falls over a period of years. These anodes also con- 
tained about 67.0 ounces of silver and 0.3 ounces of gold per ton. 

TABLE 5 a 
ANALYSIS OF ANODE COPPER TREATED AT GREAT FALLS 









Per Cent 








1929 


1930 


1931 


1932 


1933 


Copper 
Zinc 


99.290 
0008 


99 286 
0010 


99 249 
009 


99 199 
0.0010 


(3 months only) 
99 406 


Lead 


015 


020 


012 


008 




Arsenic 
Antimony 
Selenium 
Tellurium 


0.088 
072 
012 
084 


084 
073 
013 
073 


062 
075 
010 
0.089 


056 
072 
008 
069 


0.099 
0.062 


Nickel 


014 


0.022 


014 


0.010 




Bismuth 


0057 


0060 


0054 


0.008 




Iron 


0010 


0011 


0013 


0.007 




Sulfur 


0032 


0027 


0027 


0024 

















Bard well, E. S., and Lapee, R J , op cit , p 419. 



Size: 36% inches long, 28 inches wide, 2 inches thick. 
Life: 24 to 26 days produce four crops of cathodes. 
Anode scrap: 9 to 11 per cent. 
Mode of suspension: cast lugs. 



286 



ELECTROLYTIC REFINING 



Cathodes. Thirteen-hour sheets weighing 7 pounds apiece are made 
in the stripper cells. Riveted loops are attached to hold the cathode 
bar. Commercial cathodes are pulled every 6 days, and they weigh 
about 140 pounds. 

Electrolyte. The electrolyte has a specific gravity of about 1.275; 
it is maintained at a temperature of 55 to 60 C and is circulated by 
means of air lifts. Average analysis of electrolyte for two different 
years is given in Table 6. 

TABLE 6 
AVERAGE ANALYSIS OF ELECTROLYTE AT GREAT FALLS 







Grams per Liter 


Year 


Specific 




















Gravity 


Free 
Acid 


Cu 


As 


Sb 


Fe 


Ni 


Cl 


1930 


1.272 


218 


39.1 


8 2 


72 


2 9 


9 7 


027 


1932 


1.253 


213 


34.5 


8.7 


60 


2 5 


8 


028 



Bard well, E. S., and Lapee, R J , op. cit., p. 421 

Anode Mud. The anode mud produced at Great Falls is leached to 
remove the bulk of the copper and then shipped to Raritan for re- 
fining. The leached slime will assay approximately as follows: 



Cu, 
Te, 



1.5% 
15.6% 



As, 

Pb ; 



5.o%> 

4.6% 



Au, 75 oz/ton 



Sc, 

S, 



20% 

2.7% 



Ag, 13,300 oz/ton 



Purification of the Electrolyte. The method adopted for purifying 
tank-room electrolyte at Great Falls is as follows: 

Each day a certain volume of electrolyte is run off and sent to the 
purification tank, where it is boiled down to approximately 46 Baume 
(specific gravity = 1.47). The concentrated electrolyte is then sent 
to crystallizing tanks, where the bulk of the copper is crystallized out 
as copper sulfate; part of this bluestone goes to market and part is dis- 
solved and returned to the tank room (Fig. 16). The mother liquor is 
then passed through electrolytic tanks with insoluble anodes to re- 
move the last of the copper and most of the arsenic and antimony. 
Spent electrolyte from the insoluble anode cells is then either returned 
to the tank room direct or boiled down to 55 Baume (specific gravity 
1.61) , the iron and nickel salts crystallized out, and the mother liquid 
returned to the tank-room circuit as restored acid. 

Prior to 1930 all evaporating was done in lead-lined tanks provided 



GREAT FALLS 



287 




I 
I 



! 

O 



288 ELECTROLYTIC REFINING 

with lead heating coils through which 30-pound steam was passed. 
Since 1930, direct-fired evaporators have been used for this purpose. 

The basic precipitate of antimony as sulfate and oxychloride carries 
with it arsenic and bismuth (probably as basic sulfates). Solubility of 
antimony in the electrolyte is very limited, and if enough antimony is 
present most of the arsenic and bismuth will be thrown into the anode 
mud; if there is little antimony present, however, the bulk of the arsenic 
remains in the electrolyte. Table 5 shows that in 1931 and 1932 the 
arsenic content of the anodes decreased while the antimony content 
remained constant. During 1930 the purification plant was operated 
continuously in order to hold down the arsenic content of the electrolyte. 
During 1931 and 1932 the purification plant was operated intermit- 
tently, and only for short periods, to control the acid content of the 
electrolyte. In 1933 the arsenic content increased and the antimony 
content diminished so that it was necessary to resume continuous op- 
eration of the purification plant 

The 12 insoluble anode tanks which treat the mother liquor from the 
first crystallizing tanks are arranged in four cascades of three tanks 
each; antimonial lead is used for anodes and copper starting sheets 
as cathodes. The bulk of the copper is deposited in the first tank as 
impure cathodes, which are returned to the anode furnaces. The de- 
posit in the other tanks is a sludge containing most of the copper and 
arsenic together with some silver and other metals; this sludge is sent 
back to the smelting furnaces at Anaconda. 

If the tank-room solutions are low in iron and nickel the spent 
electrolyte from the insoluble-anode cells is returned directly to the 
main circuit. If nickel and iron are too high the solution is evaporated 
and a sludge of iron and nickel sulfates crystallized out; the restored 
acid then returns to the tank room. The sludge produced contains 
about 128 per cent nickel and 3.9 per cent iron; it is usually wasted, 
but it can be treated to remove the contained nickel if this becomes 
profitable. 

Montreal East. 11 The electrolytic copper refinery of Canadian 
Copper Refiners, Ltd., is located at Montreal East, Quebec. The plant 
is designed for the production of 12,500,000 pounds of refined copper 
per month, and provision is made for doubling the plant capacity 
(Fig. 17). Crude copper comes principally from the Noranda and 
Flin Flon smelters. 

Electrical Plant. Power is purchased from the Montreal Light, 
Heat, and Power, Consolidated, in the form of three-phase 60-cycle 

11 McKnight, H. S , Montreal East Plant of Canadian Copper Refiners, Limited : 
Am. Inst. Min. & Met. Eng. Trans., Vol. 106, p. 352, 1933. 



MONTREAL EAST 



289 




290 



ELECTROLYTIC REFINING 



current at 12,000 volts. The incoming voltage is reduced to 2300 volts 
and 550 volts by two banks of outdoor transformers. The power for 
the main electrolytic circuit is supplied by three motor-generator sets 
consisting of 675-kilowatt 135- to 12-volt 5000-ampere direct-current 
generators driven by 980-horsepower 2300-volt synchronous motors. 
These sets operate in parallel and supply 15,000 amperes of current to 



Bus-Bar. 



o 


3 E 


o 


o 


1 E 


o 


o 


3 E 





to o 


3 E 


/> 


o 



a E 


, 



o 

0) 


3 E 



01 


o 


3 E 


o 


Anodev 


3 E 


Cathode 


Iliui 


3 E 


lll llhl 91 






Negitlve Bm-Blf^ Positive But-Bir^ 



Solution Inlet 



Solution Overflow 

\ 



Solution Inlet 




(McKnight, Am. Inat. Mm and Met Eng Trans., Vol. 106, p. 366, 19SS) 
FIG. 18. Cell Tiers, Montreal East Refinery. 

the circuit. The power for the purification circuit is supplied by a 
125-kilowatt 25- to 3-volt 5000-ampere direct-current generator driven 
by a 225-horsepower 2300-volt synchronous motor. 

About 300 tons of copper was used in the bus system, which is about 
3500 feet long. The main runs under the working floor are made up of 
eight 10- by %-inch copper bars spaced % inch apart and in parallel. 
The tank buses are of different sizes and shapes, the largest being 17 
inches wide, 3 inches thick, and 19 feet long. The current is trans- 
mitted from bus to anode and from cathode to bus through knife-edge 



MONTREAL EAST 



291 



contacts. These contacts consist of triangular sections of copper arc- 
welded to the 3-inch busbars. 

The normal bus current is the full 15,000 amperes, and 15,000 amperes 
flows through each cell in the system, as all cells are in series in one 
electrical circuit. Cathode current density is 17 amperes per square 
foot, and the voltage drop per cell about 0.212 volt. Current density 
in the bus is approximately 375 amperes per square inch. 

Tanks. The tanks are rather large, being 16 feet 7 inches long, 3 
feet 7% inches wide, and 4 feet 1% inches deep, inside dimensions. 
They are built of reinforced concrete cast in place and lined with 6 per 
cent antimonial lead. Each tank holds 42 anod<s spaced at 4%-inch 
centers and 43 starting sheets. The cells are supported independently 
of the working floor by concrete columns of such an elevation that the 
tops of the cells are 18 inches above the working floor. The columns 
are capped with a glass plate 1% inches thick, a rubber sheet, and a 
lead shield to insulate the cells from the ground. The cell bottoms 
are 7 feet above the basement floor to provide ample space for inspec- 
tion and headroom for solution lines and slimes launders. 

There are 468 tanks 432 commercial and 36 stripper tanks 
arranged in tiers of 9 cells each and grouped in sections of two tiers or 
18 cells. There are 26 such tiers, and all the cells are connected in 
series. The anodes and cathodes in any one tier are directly connected 
by means of " Baltimore grooves " cast in the top of one lug of each 
anode and accommodating the adjacent cathode bar. The electrical 
connections in a section are shown in Figure 18. 

Anodes. About 60 per cent of the anodes come from the Noranda 
smelter already cast. The remainder are cast at the plant from Hud- 
son Bay (Flin Flon) blister. The two lots have slightly different 
compositions, as may be seen in Table 7. 

TABLE 7 a 
TYPICAL ANALYSES OF ANODES AT MONTREAL EAST* 





Cu 


Ag 


Au 


As 


Sb 


Se-f-Te 


Pb 


Ni 


Fe 


S 


Noranda 
Domestic 


99 41 
99 49 


17 09 
31 40 


8 058 
4 577 


0024 
0126 


0023 
0073 


264 
0.183 


0.0016 
0.0236 


0293 
0.0143 


0.0065 
0.0083 


0.0140 
0.0155 



tt McKmght, II. S , op cit. p 355. 

* Au and AS in ounces per ton, all others in percentage. 

Weight, 700 pounds. 

Size, 36 by 36 inches on face; 1$ inches thick. 

Life, 33 days produce 2 crops of cathodes. 

Anode scrap, 14 per cent. 

Mode of suspension: cast lugs; Baltimore groove. 



292 



ELECTROLYTIC REFINING 



Cathodes. The starting sheets are 37 l / 2 inches square slightly 
larger than the anode surface; these are deposited in stripper cells 
using anodes 'weighing 770 pounds (slightly larger than the anodes in 
the commercial cells). Each cathode weighs about 300 pounds and is 
removed at the end of 16 days. Two crops of cathodes are made from 
each anode, and when the first set of cathodes is removed the anode 
spacing is closed up to 4 inches and five additional anodes are added 
to each cell. Starting sheets are supported by a single large loop at 
the center. 

Electrolyte. The electrolyte flows by gravity from head tanks 
through the electrolytic cell to sump tanks in the pump bay located 
below the basement floor. It is then elevated by centrifugal pumps 
to the head tanks, where it is heated by steam coils. The temperature 
is maintained at 140 F (60 C) and the flow through each cell is 4.5 
to 5 gallons per minute. Commercial and stripper , electrolytes are 
separate. 

Circulation within the cell is from bottom to top (Fig. 18). The 
inlet to each cell is of rubber to prevent loss of current through the 
solution lines. The overflow line is broken by an air gap in a lead 
" boot." 



TABLE 8 a 
TYPICAL ANALYSIS OF ELECTROLYTES AT MONTREAL EAST 



Electrolyte 


Temp 


Deg B6 

at60F 


Free 
H 2 S0 4 

(%) 


Cu 


Specific 
Gravity 


Cl 


Stripper 
Commercial 


140 
140 


22.50 
27.30 


12.96 
16 60 


3.16 
3.20 


1.19 
1.24 


0.00075 
0.00019 



McKmght, H. S , op cit , p 357. 

Anode Mud. Slimes are removed at the end of the first cathode run, 
but the anodes are not moved. At the end of the second run the anode 
scrap is removed and the tanks are completely emptied and cleaned. 
The clear electrolyte is pumped from the cell to within a few inches of 
the bottom; then the plug is pulled and the slime and remaining 
electrolyte are conducted through lead-lined launders to tanks from 
which they are pumped to storage tanks in the silver refinery. 

These slime-bearing solutions are pumped to a Dorr thickener, and 
the overflow solutions are returned through a settler to the tank room. 
The thickened slimes are filtered on an Oliver filter, roasted for l l / 2 
hours at 500 to 650 F (260 to 343 C), and then leached for 3 hours 



ONTARIO REFINING COMPANY 



293 



in 20 per cent acid solution made up chiefly of foul electrolyte from 
the tank room. The leached slimes are washed and filtered and de- 
livered to the dore furnace. 

TABLE 9 a 
TYPICAL ANALYSER OF SLIMES AT MONTREAL EAST, IN PER CENT 



Slimes 


Cu 


Ag 


Au 


As 


Sb 


Se 


Tc 


Pb 


Ni 


Fe 


S0 4 


SiO 2 


Raw 


45 


8 


2 5 


120 


222 


24 61 


3 77 


1 91 


037 


338 


6 21 


2 18 


Roasted 


42 


7 5 


2 3 


083 


107 


19 17 


3.18 


3.77 


014 


358 


5 61 


2 17 


Leached 


3 8 


31 


9 3 


076 


242 


24.62 


2.43 


10 16 


0.006 


1 030 


5.38 


6 56 



McKmght, H S , op cit , p 365 

The decanted leach solutions are filtered and passed through eight 
electrolyte cells employing lead anodes and lead cathodes. Practically 
all of the coppeV and selenium is removed in these cells in the form of a 
copper-selemunAsludge. These solutions are then cemented with iron 
to remove the ikst traces of copper and are then discarded. The 
copper-selenium sludges are stored for possible future treatment to 
remove the selemur 

The leached slimes, containing about 25 per cent moisture, are 
treated in the dore iurnace with the necessary fluxes of soda and niter. 
The first " scoria " qr slags from the dore furnace are returned to the 
anode furnace for recovery of gold, silver, and copper contents. The 
soda and niter slags ar^ leached with water, and the residue is returned 
to the dore furnace. 

The dore bullion is castS^to anodes and refined in 10 Moebius cells. 
Cathode silver is melted and a$t into 1000-ounce bars of a fineness of 
999 + . The gold slime is boilech-^th concentrated sulfuric acid to 
remove silver and is then cast directly iiTtrarrthat are 992 -f fine. 

Gases from the dore furnace and roasting furnace pass through cool- 
ing flues, a scrubber system, and a Cottrell treater. The Cottrell 
treater is of the pipe type and is made entirely of lead and lead-covered 
steel. 

Ontario Refining Company. 12 The electrolytic copper refinery at 
Copper Cliff, Ontario, treats the crude copper from the smelting of 
the nickel-copper ores of the Sudbury district. The refinery has a 
capacity of 20,000,000 pounds of refined copper per month. The 
anodes contain larger amounts of nickel than that found at most 
refineries. 

12 Benard, Frederic, Electrolytic Copper Refinery of Ontario Refining Com- 
pany, Ltd., at Copper Clift, Ontario: Am. Inst. Mm. & Met. Eng. Trans., Vol. 106, 
p. 369, 1933. 



294 



ELECTROLYTIC REFINING 



Electrical Plant. Part of the power required is purchased, and part 
of it is generated by a 2500 kva turbogenerator at 2300 volts; the 
turbogenerator is operated by steam from waste-heat boilers and a 
pulverized coal-fired auxiliary boiler. Power is purchased at 30,000 
volts which is stepped down to 2300 volts in a bank of three 2000- 
kilovolt-ampere transformers, with a fourth as a spare. Three motor- 
generator sets supply the electrolytic load; each consists of one 



Transformer 



To I.N. Co, Smelter 



Add Storage 

Gas Generating It 

Room || 

" GaTstoTigi j 




1 Lumber Storage 

2 Carpenter Shop 

3 Pipe Shop 

4 Electrician Shop 

5 Locker Room 

6 Toilet 

7 Office 

8 Tool Room 

9 Machine Shop 

10 BUcksmlth & Plate Shop 16 Anode Furnaci 

11 Inspection Conveyor 17 Pump Plt-Aux. 

12 Boah Conveyor Boiler Above 

13 Wire Bar Casting Wheel 18 Shop and Office 

14 Wire Bar Furnace 19 Compressor House 

15 Anode Catting Wheel 20 Wash and Locker Room 



(Benard, Am. Inat Mm and Met. Eng Trans , Vol 1O6, p 371, 1933) 

FIG. 19. General Plan of Refinery of Ontario Refining Company. 

2880-horsepower 500-revolutions per minute synchronous motor driv- 
ing two 6000-ampere 80- to 160-volt direct-current generators, and 
each set is rated at 2500 kilovolt-amperes. The third set is a spare 
and may be used interchangeably with the other two. 

Current density is about 15 amperes per square foot with a voltage 
drop of 0.180 to 0.200 volt per cell. Current efficiency is 97 to 98 
per cent. 

Tanks. The tank house contains 1230 concrete tanks lined with 
6 per cent antimonial lead. These are laid out in 32 sections of 38 
tanks each, with two liberator units of six tanks each and two full-sized 
experimental tanks. The tanks are 11 feet 3 inches long, 3 feet 6 inches 
wide, and 3 feet 9% inches deep, inside dimensions. The tanks are 



ONTARIO REFINING COMPANY 



295 



supported by concrete piers with a 2-inch slab of vitrified tile and a 
lead cap between the pier and the tank sill; the tile is for electrical 
insulation from the ground and the lead cap for the protection of the 
pier from drip. There is 9% feet of headroom between the basement 
floor and the tank bottoms. 



30 Ventilator 
every second Bay 




(Benard, Am Inst Min and Met Eng Trans , Vol 106, p. 575, 19SS) 

FIG. 20. Cross-Section Through Tank House, Ontario Refining Company. 

The tanks are arranged in two electrical circuits, which cross at 
right angles the two solution circuits. A current of 10,500 amperes 
passes through each electrical circuit. The tanks are connected by 
the Walker multiple system of contacts. Each commercial tank con- 
tains 38 anodes and 39 cathodes; stripper tanks contain 31 anodes, 
30 mother blanks, and 2 end sheets. There are 76 stripper tanks 
(2 sections) and 1140 commercial tanks. 

Anodes. The anodes are cast from incoming blister cakes, in two 
36-foot Walker casting wheels. 

Weight, 530 pounds. 

Size, 36 by 36 inches; 1}4 inches thick. 

Life, 28 days produce two crops of cathodes. 

Anode scrap, 12 per cent. 

Mode of suspension, cast lugs. 



296 ELECTROLYTIC REFINING 

Cathodes. Starting sheets are stripped every 24 hours and weigh 
10 pounds each. Loops 4 inches wide of the same material are 
punched into them by electrically driven double-punch machines. 
Finished cathodes weigh 240 pounds and are pulled at the end of 
14 days. 

Electrolyte. The electrolyte is held at about 3 per cent copper 
and 13 per cent free acid. The solution is circulated at the rate of 
2% to 3 gallons per minute and is heated to 150 F (65 C). Glue, 
oil, and bindarine are added in small quantities to improve the tough- 
ness and smoothness of the cathode deposit. Circulation through the 
cells is from bottom to top. 

Anode Mud. The analysis of the anode slimes are given below. 
The gold and silver assays are omitted because these are rather 
variable owing to the variation in custom material treated. 

Per Cent Per Cent 



Copper 24 70 

Nickel 19 80 

Selenium 15 03 

Tellurium 3 61 

Iron 40 



Silica 0.18 

Lead 1.51 

Arsenic 24 

Antimony 32 

S0 4 5 48 



Note the abnormally high nickel content. Every effort is made to 
throw as much of the nickel as possible into the electrolyte instead of 
the slime. 

Slimes are roasted and leached to remove copper and nickel; this 
is done by mixing the raw slimes with concentrated sulfuric acid and 
roasting in a reverberatory furnace with a sectional cast iron bottom. 
The roasted slimes are given one leach with 10 per cent sulfuric acid 
and one water leach. The nickel and copper sulfate solution that 
results is treated with copper sludge to precipitate any selenium, 
tellurium, or silver, and returned to the tank house. The precipitate 
joins the regular slimes. 

The treated slimes are refined to dore metal in an oil-fired reverbera- 
tory furnace which treats about 14,000 pounds of treated slime con- 
taining 20 per cent moisture. The slimes are mixed with a flux of 
1 per cent sand and 2 per cent fluorspar, and slag is skimmed before 
the furnace is fully charged. After the charge is melted and the 
scoria slag skimmed, the matte is refined to dore, using soda ash and 
niter. The scoria slag is returned to the anode furnaces. Soda-niter 
slags are leached with water ; the residue is returned to the dor6 furnace, 
and the leach liquor is pumped to the selenium plant. 

The dore bullion is cast into anodes and refined in 40 Balbach 



ONTARIO REFINING COMPANY 



297 



(Thum) cells. The gold slime is digested with aqua regia and the gold 
reprecipitated with ferrous chloride. This gold is then melted, cast 
into anodes, and refined in 6 Wohlwill cells. The solution remaining 
from the aqua regia leach is cemented on scrap iron to precipitate the 
platinum metals, and the platinum residues are shipped to a refinery 
in Acton, England. 

The refinery also includes a plant for the recovery of selenium and 
tellurium from furnace slags and flue dusts. 



(INCOMING BLISTER I 

I COPPER 1 



FURN, 
ANODI 
WIRE BAR 



INTERNATIONAL 
NICKEL CO. 
SMELTER 




n 


* 
SCALES 1 








f 




y J DRILL ROOM) 






^ SLAG nrr 


t . 








IMPURE 


[ 


ANODES"] 


O 




SOLUTION 1 TANK HOUSE s 


LIMES CONTAINING Au, Ag Pt. Pd, and Se 









i | CATHODE'S"] 


ISILVER REFINERY] ' 


A 1 


, 


< OLU T!?.fV.|sETTLIN 




ACID RECOVERY 
PLANT 


S 

X 




1 LIBERATOR 1 
| TANKS | 


1 WIRE-BAR 1 I SHEAR 
v 1 FURNACE | ' -J 


[ROASTER FURNACE] 


ATORl 
'ER 1 




J 5 


Z \ ^ 1 JLEACH TANKS] 


DE COPPERIZED] g, 

SOLUTION 1 r^ 

r ..,. i. ., | WH 


8 1 MARKET 

ING 1 
EEL 1 







.,., 

' 1 CENTRIFUGE| 

* 1 


CES, | EVAP 
AND 


* WIRE 


BARS 


DORE METALi . | i LOW GRADE 



IPARTING CELLS] 

FINE SILVERfgOLD MUD t | 



SILVER CHLORIp.EiSO.LUTION CONTAINING 
' f ^ ' | Au, Pt, and "0 

' EDUCTION] (PRECIPITATING TANKS] 

GOLD SAND | SOLUTION Pt. Pd 




(Benard, Am Inst Min and Met Eng Trans , Vol 106, p. 373, 1933) 

FIG. 21. Flowsheet, Ontario Refining Company. 

Purification of the Electrolyte. The principal impurity is nickel, 
and to remove this impurity a given amount of electrolyte must be 
removed from the tank house each day. Nickel markedly increases 
the electrical resistance of the electrolyte. By removing the nickel, the 
other impurities are easily held below the required limits. 

The first step is to pass the solution through Pyne-Green segregating 
tanks, which have a bottom inflow, bottom outflow, and a small orifice 
near the top through which passes less than 10 per cent of the 
total flow. By means of the restricted circulation at the top, a segre- 



298 ELECTROLYTIC REFINING 

gation takes place, and the upper layers of solution are depleted of 
their copper ions, the copper content being reduced from 3 to 1 per cent 
and the nickel and acid content being somewhat increased. This 
" segregated " solution is then delivered by automatic air lifts to the 
acid recovery plant. 

The segregated solution is first passed through a nest of 10 tanks 
carrying 5000 amperes and using 25 lead anodes and 24 regular 
copper starting sheets. These tanks reduce the copper content from 
12 grams per liter (1 per cent) to 8 grams per liter and produce com- 
mercial cathodes' which go to the wirebar furnaces. 

The same solution is then passed to 10 similar tanks which reduce 
the copper to 0.2 gram per liter and produce impure cathodes which 
go either to the anode furnaces or to the silver refinery for cementa- 
tion purposes. 

The decopperized solution is then evaporated from 23 Baume to 
about 60 Baume; nickel and iron residues are crystallized and pre- 
cipitated, and the clear acid (54 Baume) is returned to the tank house. 

Phelps Dodge Refinery. 13 The refinery of the Nichols Copper 
Company at Laurel Hill, New York (now known as the Phelps Dodge 
Refinery), utilizes the Nichols system of series refining using cast 
anodes, and we shall present a description of this plant to illustrate 
the series system. 

One tank room of the Baltimore plant of the American Smelting and 
Refining Company utilizes the Hayden series system in which the 
impure electrodes are rolled rather than cast. Certain impurities make 
it difficult to roll the copper properly, and the Hayden system is thus 
restricted to the refining of lean high-grade bullion. This restriction 
does not apply to the Nichols system. The Laurel Hill plant treats 
a wide variety of crude copper which is of the same type as that treated 
in refineries using the multiple system. This plant has a capacity of 
about 35,000,000 pounds of copper per month. 

Tanks. The tanks used are either of wood or concrete construction 
and measure 16 feet long, 5 feet 4 inches wide, and 5 feet 2 inches deep, 
inside dimensions. The bottom lining is made of blown oil, asphalt, 
silica sand, and powdered silica; the side lining is a %-inch layer of 
blown oil mopped on in successive layers and burned on with a hot 
iron. The tank is capped with stationary wooden spacing blocks on 
each side forming slots into which the 102 iron suspension bars can 
fit; each bar carries 5 anodes (Fig. 23), and this group of 5 electrodes 
is known as a cell. Each tank, when loaded, contains 510 regular 

13 Harloff, C. S., and Johnson, H. F., The Nichols System of Electrolytic Cop- 
per Refining- Am. Inst. Min. & Met. Trans., Vol. 106, p. 398, 1933. 



PHELPS DODGE REFINERY 



299 



anodes in 102 cells; the charge amounts to 56,000 pounds of metal 
and 22,000 pounds of electrolyte per tank. The tanks are electrically 
connected in groups of about 44 in two parallel lines of 22 each (all 
groups are not uniform), and there are 417 tanks in the tank house. 
All are " commercial " cells, of course; starting sheets and " stripper " 
cells are not needed. 

Tanks are loaded and unloaded by 5-ton cranes which will pick up 
17 bars (85 anodes) at a load, thus requiring six loads per tank. 



102 cells 
per tank 
1 Cell- 
5 Plates 




(Harloff and Johnson, Am. Inst. Mm and Met Eng Trans., Vol 106, p. 401, 1933) 

FIG. 22. Plan View of Loaded Series Tank Showing Set-up, Spacing Blocks, and 

Circulation Inlet. 

A copper busbar 2% inches by % inch connected by a copper cable 
to the negative bus of the preceding tank serves to bring current into 
the tank at the positive end. Five 150-pound anodes are hung from 
the positive busbar, and here the current is introduced into the tank. 
At the negative end is a similar busbar from which are suspended five 
45-pound depositing cathodes. The remaining electrodes in the tank 
are suspended from iron bars by means of copper links (Fig. 23). 
These electrodes weigh about 110 pounds each. Each of these 110- 



300 



ELECTROLYTIC REFINING 



pound electrodes serves as a bipolar or intermediate electrode, receiving 
a deposit of pure copper on one side while impure copper is being 
dissolved from the opposite side. 

Electrodes. The electrodes are called anodes before they are put 
into the tank and cathodes after they are removed. The 110-pound 
anodes, which are 56 by 12 inches by % 6 inch with two ears forming the 
top part, are carefully cast by hand. The ears are subsequently 
punched out to form suspension lugs (Fig. 23). These anodes are 
suspended from steel hanger bars by acid-resistant cast links of copper 
alloy. 




(Harloff and Johnson, Am. Inst Mm. and Met Eng Trans , Vol 106, p 402, 1933) 

FIG. 23. Elevation of Series Electrolytic Tank Showing Spacing of Plates, 
Circulation Glass, and Overflow. 

After the casting and partial cooling, the warm anodes are punched, 
carefully straightened, assembled on the suspension bars, and spray- 
painted with a hot (180 F) 25 per cent solution of sodium resinate on 
the side which is to receive the pure copper deposit. This forms a 
cleavage zone that facilitates the removal of anode scrap without 
materially increasing the electrical resistance of the depositing surface. 



PHELPS DODGE REFINERY 301 



Anodes are spaced l^Ke inches center to center with a 1-inch 
clearance on the side of the tank and % inch between adjacent anodes 
on the same suspension bar. New anodes are loaded into an empty 
tank and carefully adjusted so that all anodes are suspended properly. 
Electrolyte is then fed into the tank by a 2% -inch hose from a valve 
on the main circulating line until about 6 inches of the anodes is 
submerged. The cut-out bar is then disconnected, and current flows 
through the tank while the remainder of the electrolyte is added. This 
procedure is adopted in order to put an initial cathode deposit on the 
resinous paint as soon as possible; if this is not done the paint will 
dissolve in the hot electrolyte and the cathode deposit will cling to the 
anode so tightly that the anode scrap cannot be removed. 

The current density will range from 13 to 26 amperes per square foot, 
depending upon the production desired; accordingly the tank cycle 
will range from 13 to 34 days. The cycle is so adjusted that when 
the cathodes are removed only a thin sheet of the original anode re- 
mains. The anode scrap is then stripped by hand from the cathode 
deposit and returned to the anode furnaces. Anode scrap amounts 
to 6 or 7 per cent of the original anode weight. 

Electrolyte. The copper content of the electrolyte is maintained at 
2.75 to 285 per cent copper and 17.5 to 18.5 per cent free acid. Part 
of the Cu 2 O in the anodes is chemically dissolved in the electrolyte to 
replenish the copper loss, and by careful regulation it is possible to 
avoid the use of insoluble-anode cells to plate out excess copper from 
the electrolyte. A certain amount of the electrolyte is withdrawn and 
sent to the copper sulfate plant, and this keeps the impurities within 
the proper limits. 

Cathodes of the desired purity have been produced by using an 
electrolyte containing up to 1.0 per cent nickel, 0.35 per cent arsenic, 
06 per cent antimony, and 0.05 per cent iron. Sodium chloride is 
added daily to keep the chlorine content of the solution at 0003 to 
0005 per cent. Glue and Goulac are added to improve the physical 
characteristics of the cathode deposit; about an ounce of glue and a 
pound of Goulac are added for every 50 tons of cathodes deposited. 

The electrolyte is heated to 122 F (50 C) and is circulated through 
the tanks at about 3 5 gallons per minute. A glass tube introduces 
the electrolyte at a point about 5 to 8 inches above the tank bottom, 
and the solution overflows at the top of the tank. 

It is necessary to add sulfuric acid and water to the electrolyte to 
replace the acid lost by chemical action and the water lost by evapora- 
tion. Part of the replacement water comes from that used to wash 
the cathode deposits. 



302 ELECTROLYTIC REFINING 

Anode Mud. The anode mud is cleaned out of each cell whenever 
the cathodes are withdrawn. This will amount to about 0.8 per cent 
of the anode weight, depending on the purity of the anodes. This is 
screened and washed, roasted, and leached to remove the copper. The 
residue is shipped to another plant for treatment. The leach liquor is 
treated in the copper sulfate plant. 



CHAPTER IX 

HYDROMETALLURGY 

INTRODUCTION 

In general the hydrometallurgy of copper refers to those processes 
by which copper-bearing material is leached with a solvent to dissolve 
the copper, the solution being then separated from the residual solid or 
tailing and the copper precipitated from it. Solutions used have always 
been aqueous solutions, as the name hydrometallurgy suggests. 

Of all the methods employed for recovering copper, leaching is one 
of the oldest, and it is the most complex of all, judging from the large 
number of experimental, semi-commercial, and commercial processes 
that have been developed. These include a wide variety of solvents 
and precipitation methods and a number of different methods of purify- 
ing solutions and mechanical handling of materials. 1 ' 2 Leaching 
methods have been used on ores (both mined ore and ore still in place) 
on concentrates, calcines, mattes, and other products. 

We shall find it necessary to confine our discussion largely to those 
methods which are in commercial use today. It is to be noted that 
leaching processes depend upon the nature of the ore being treated in 
an even greater degree than pyrometallurgical treatment. Except for 
relatively minor differences, it is possible to discuss copper roasting, 
smelting, converting, and refining in a general way and still not deviate 
too much from the practice found in special cases. Hydrometallurgical 
processes, however, differ from one another to such an extent that it 
is difficult to give a general discussion that will include all processes. 

Crude ore in which the copper occurs as oxidized minerals (or oxidized 
minerals plus some sulfides) and some native copper tailings and ores 
constitute the two large classes of materials which are subjected to 
leaching. The leaching agents are generally either (1) sulfuric acid 
or (2) ammonia. The most widely used method of precipitating the 
dissolved copper is by electrolysis ; the only other method of importance 
is cementation on scrap iron. 

1 Hofman, H. 0, and Hayward, C. R., Metallurgy of Copper: McGraw-Hill 
Book Co., New York, 1924. 

2 Greonawalt, W. E., The Hydrometallurgy of Copper: McGraw-Hill Book Co., 
New York, 1912. 

30;* 



304 HYDROMETALLURGY 

We shall consider the question of electrolysis later in the chapter, 
but' at this point we shall mention the fact that this process differs 
from the electrolytic refining of copper in many respects. The recovery 
of copper by electrolysis is often called elcctrowmnmg as distinguished 
from electrorefimng discussed in the preceding chapter. To the com- 
bination of leaching and electrolysis the term clcctro-hydrometallurgy 
is often applied. 

General Considerations. In any leaching operations there are three 
important steps: 

1. Bringing the solvent in contact with the material to be leached 
to permit dissolution of the metal. 

2. Separating the charged or pregnant solution from the solid 
residue. 

3. Precipitating the metal from the solution. As a general rule 
these follow in the order indicated, but sometimes the order of the 
last two are reversed, i.e., copper is precipitated while still in the 
ore-water mixture and is then separated from the pulp by flotation. 

The solvent used for leaching must have the following characteristics: 

1. It must be cheap and available in adequate quantities. 

2. Jt must have a selective action, i.e., it should attack the ore 
minerals but not the gangue minerals. 

3. In general it mu^t be effective when used in cold dilute solutions. 

4. If possible it should be regenerated by the precipitation operation. 
These characteristics are modified by the nature of the material 

being leached. It is possible to treat roasted anode mud by boiling 
in concentrated sulfuric acid, but this method would hardly be ap- 
plicable to the treatment of a 1 per cent copper ore this would per- 
haps be treated by a prolonged leaching with cold dilute acid. Aside 
from the expense involved, the boiling sulfuric acid would probably not 
show a selective action with reference to either the ore or the leaching 
tanks. 

The leaching method employed will also depend upon the nature of 
the ore involved, and in general there are two: 

1. Percolation or sand leaching. 

2. Agitation or slime leaching. 

The terms sand and slime leaching are used principally in the 
cyanide leaching of gold and silver ores; sand (usually fairly coarse 
material) will allow solution to flow or percolate freely through the 
interstices, but slime (fine material) will pack in a vat or tank and 
impede the circulation of the liquid. Sands may be leached by simply 
allowing the liquid to stand on the material for a given time, but slimes 
require that the slime-liquid pulp be kept in agitation during the 



ORES SUITABLE FOR LEACHING 305 

leaching period to prevent the solids from settling and packing. We 
shall find applications of both methods in the leaching of copper ores. 
The precipitation method used may be either (1) electrolytic or 
(2) chemical. Electrolytic precipitation is expensive but it yields a 
pure copper equal in grade to electrorefined copper, and it regenerates 
the sulfuric acid used as a solvent. Electrolytic precipitation is not 
used with ammonia leaching. Chemical precipitation is usually cheaper 
than electrolysis and can be used on some solutions which contain too 
little copper for satisfactory electrolysis. It produces an impure pre- 
cipitate which must usually undergo further smelting and refining. 

ORES SUITABLE FOR LEACHING 

Hydrometallurgy has many advantages over pyrometallurgy it 
generally requires no fuel or expensive furnaces, and often the prin- 
cipal expense is for the reagents used in the solvent. The principal 
disadvantage is simply the fact that many copper ores do not respond 
to leaching methods well enough so that the hydrometallurgical proc- 
esses can compete with pyromctallurgy. Today only a rather small 
class of ores is being treated by leaching, but eventually processes may 
be developed which will widen the scope of leaching methods as applied 
to copper ores. 

Commercial leaching plants operating today treat ores in which the 
copper minerals are (1) oxidized (soluble in water or dilute sulfuric 
acid), (2) partly oxidized and partly sulfidc (sulfides are soluble in 
acid ferric sulfate), or (3) native copper or carbonate minerals (soluble 
in ammoniacal solutions). The choice of solvent to be used and 
purification and precipitation methods depends on a number of factors; 
among the most important are: 

1. Solubility of the Ore Minerals. The only practical solvent for 
copper is an ammoniacal solution; ammonia can also be used to 
dissolve carbonates. Chalcanthite (CuSO^SHoO) is soluble in water; 
other oxidized copper minerals are soluble in dilute acid. Sulfide 
minerals can be dissolved in acidified ferric sulfate solution. 

2. Solubility of the Gangue Minerals. The gangue must be only 
slightly soluble in the solvent to avoid the use of excess reagent and to 
prevent serious fouling of the solution. Thus, for example, ores with 
a gangue of dolomite or limestone must be leached with ammonia 
because these substances are very soluble in acids. 

3. Nature of the Copper Minerals. Copper in the form of chalcan- 
thite dissolves without consumption of acid and actually generates 
free acid in the circuit when electrolytic precipitation is used. Oxidized 



306 HYDROMETALLURGY 

minerals dissolve in sulfuric acid, and the acid consumed is regenerated 
by electrolytic precipitation. 

4. Soluble Impurities in Gangue and Ore Minerals. When copper 
is to be precipitated electrolytically the impurities in the solution must 
be carefully controlled by means of a purification system. The 
impurities which are most harmful to the electrolysis are: 

(a) Chlorine Ions. Chlorine ions are usually found when the ore 
contains atacamite (CuCl 2 '3Cu(OH) 2 ). 

(b) Nitrate Ions. Nitrate ions are formed from soluble nitrates in 
the ore. 

(c) Iron. The concentration of both ferric and ferrous iron is 
important. 

(d) Molybdenum. 

5. The Per Cent of (a) Total Copper, (b) Sulfide Copper and (c) 
" Acid Soluble " Copper in the Ore. 

SOLVENTS 

With the exception of certain solvents used in special processes, e.g., 
using strong sulfuric acid to leach copper from roasted anode mud, 
the important leaching agents used on copper ore are water, sulfuric 
acid, and ammonia; ferric sulfate is of some importance, but it is used 
with sulfuric acid, and the acid is the principal solvent. 

Sulfuric Acid. Acid used for leaching will usually contain 25 to 70 
grams of free acid per liter of solution, although this will vary, depend- 
ing upon the stage of the operation. A given charge of ore will usually 
be treated with a succession of solvents of different strength. The 
chemical action of sulfuric acid upon the common oxidized minerals is 
given below: 3 

Azurite and Malachite. Both azurite and malachite dissolve readily 
in dilute acid according to the equations 

Azurite: 

Cu 3 (OH) 2 (C0 3 )2 + 3H 2 S0 4 -> 3CuS0 4 + 2C0 2 + 4H 2 
Malachite: 

Cu 2 (OH) 2 C0 3 + 2H 2 S0 4 -> 2CuS0 4 + C0 2 + 3H 2 

Tenorite (melaconite) . Tenorite (CuO) dissolves readily in dilute 
acid according to the equation 

CuO + H 2 S0 4 -> CuS0 4 + H 2 

3 Sullivan, J. D., Chemical and Physical Features of Copper Leaching: Am. Inst. 
Min. & Met. Eng. Trans., Vol. 106, p. 515, 1933. 



SULFURIC ACID 307 

Cuprite. Cuprite (Cu 2 0) does not dissolve as readily as tenorite, 
and in a solution containing only sulfuric acid only half of the copper 
goes into solution: 

Cu 2 O + H 2 S0 4 -> CuS0 4 + Cu + H 2 O 

If the solution contains an oxidizing agent, the precipitated copper 
will oxidize and then dissolve as CuS0 4 . Oxygen from the air or 
dissolved ferric sulfate will serve to oxidize the copper and permijr 
complete dissolution; ferric sulfate is much more active than oxygen 
from the air dissolved in the solution. 

Chrysocolla. The mineral chrysocolla is easily dissolved in sulfuric 
acid with the formation of CuS0 4 and liberation of silica: 

CuSi0 3 *2H 2 O + H 2 S0 4 -> CuS0 4 + Si0 2 + 3H 2 

The term " chrysocolla " is properly applied only to the mineral whose 
composition is given above. The name " chrysocolla " is often given 
to all copper silicates, and it is stated that some forms of chrysocolla 
are soluble in sulfuric acid whereas other forms are not. True chryso- 
colla is readily soluble; dioptase, another copper silicate, dissolves 
much more slowly than chrysocolla. Other minor silicates of copper 
are bisbeeite, cornuite (the amorphous equivalent of crystalline chryso- 
colla), plancheite, and shattuckite] the chemical properties of these 
minerals are not definitely known. 

Brochantite. Brochantite is a basic sulfate, the most important 
mineral at Chuquicamata, Chile, and is readily soluble in sulfuric acid. 

Cu 4 (OH) 6 SO 4 + 3H 2 S0 4 -> 4CuS0 4 + 6H 2 O 

The sulfate chalcanthite (CuS0 4 *5H 2 0) is soluble in water. 

Atacamite. The basic chloride, atacamite, dissolves in sulfuric acid 
according to the reaction 

Cu 2 Cl(OH) 3 + 2H 2 S0 4 -> 2CuSO 4 + HC1 + 3H 2 

Note that this dissolution adds chlorine ions to the solution. 

These are the important reactions for the dissolution of copper 
minerals in sulfuric acid, and they all involve simple double decompo- 
sition with the formation of soluble CuS0 4 . Other minerals can be 
dissolved by the combined action of sulfuric acid plus an oxidizing 
agent; these we shall consider later. 

In addition to the copper minerals, the various oxides, silicates, and 
carbonates in the gangue are attacked by the acid to form soluble 
sulfates of iron, aluminum, magnesium, etc. The amount and compo- 
sition of these dissolved impurities depend upon the nature of the 



308 HYDROMETALLURGY 

ore; they result in increased consumption of acid and they contaminate 
the solution. 

Water. The only copper mineral soluble in water is chalcanthite, 
CuSO 4 '5H 2 0. This is not a common mineral in natural ores, but it is 
formed by slow oxidation and weathering of sulfide ores, and hence 
is quite important. 

When copper sulfate solutions are electrolyzed using an insoluble 
anode, one mol of H 2 S0 4 is generated for each mol of CuS0 4 decom- 
posed. Hence such sulfates as chalcanthite and brochantite are acid- 
forming minerals because less acid is required to dissolve them (none 
for chalcanthite) than is generated in the electrolytic cells when the 
dissolved copper sulfate is decomposed. 

Acid Ferric Sulfate. The combination of H 2 S0 4 plus an oxidizing 
agent will dissolve many copper minerals which are not soluble in 
acid alone. The most common oxidizing agent is ferric sulfate, and 
this is not added as a reagent but is formed by the dissolution of 
iron-bearing minerals in the ore. The essential reaction involved is 
the oxidation of material by the ferric sulfate which is itself reduced 
to ferrous sulfate. 

Cuprite. Sulfuric acid alone, as we have noted, attacks cuprite to 
form Cu and CuS0 4 , thus putting only half of the copper in a soluble 
form. Ferric sulfate will attack the metallic copper: 

Cu + Fe 2 (S0 4 ) 3 -> CuS0 4 + 2FeSO 4 / 
Ferric sulfate also assists the reaction 

Cu 2 O + H 2 SO 4 -> Cu + CuS0 4 + H 2 O 

because the precipitated copper is dissolved before it can form an 
impervious coating on the outside of the cuprite particle. 

Chalcocite. Chalcocite will dissolve in ferric sulfate according to 
the reaction 

Cu 2 S + 2Fe 2 (SO 4 ) 3 -* 2CuSO 4 + 4FeS0 4 + S 

Chalcocite dissolves in two distinct stages; at first it dissolves rather 
rapidly until about half the copper is gone (atomic ratio 
Cu : S = 0.9 : 1) after which the dissolution of copper proceeds more 
slowly, indicating that the previous reaction may take place in two 
stages: 4 

Cu 2 S + Fe 2 (S0 4 ) 3 -> CuS + CuS0 4 + 2FeS0 4 
CuS + Fe 2 (SO 4 ) 3 -> CuS0 4 + 2FeS0 4 + S 
4 Sullivan, J. D., op. cit., p. 522. 



AMMONIA 309 

The sulfur liberated when chalcocite is leached remains behind as a 
solid; completely leached chalcocite appears as porous lumps of sulfur 
of about the same size and shape as the original chalcocite particles, 
but these will dissolve almost completely in carbon disulfide (a solvent 
for elemental sulfur). Aerated sulfuric acid has only a very slight 
solvent effect on chalcocite, and it appears that the oxidation of chalco- 
cite by " weathering " or natural agencies is due to the presence of 
ferric salts. 5 

Chalcocite is the most significant sulfide as far as leaching operations 
are concerned because it is associated with the oxidized copper minerals 
in many of the commercially important deposits. 

Bornite. Bornite (Cu 5 FeS 4 ) responds to an acid ferric sulfate leach 
in much the same way as chalcocite, and the copper is dissolved at 
about the same rate. 

Covellite. The reaction given for the dissolution of this mineral is 
the same as the second stage in the dissolution of chalcocite: 

CuS + Fe 2 (S0 4 ) 3 -> CuSO 4 + 2FeS0 4 + S 

Even though this reaction represents the slower stage in the dissolution 
of chalcocite, the " artificial covellite " dissolves much more rapidly 
than natural covellite. Covellite dissolves in acid ferric sulfate much 
less readily than chalcocite or bornite. 

Chalcopyrite. Under ordinary conditions this mineral is prac- 
tically insoluble in acid ferric sulfate solutions unless the ore is first 
roasted or hot solutions are used. It appears that when dissolution 
does take place there is no selective dissolution of any part of the 
molecule ; part of the sulfur is oxidized and part is liberated according 
to the reactions 

CuFeS 2 + 2Fe 2 (S0 4 ) 3 - CuS0 4 + 5FeSO 4 + 2S 
CuFeS 2 + 2Fe 2 (SO 4 ) 3 + 2H 2 O + 3O 2 -> CuS0 4 + 5FeSO 4 + 2H 2 S0 4 

Tetrahedrite, Tennantitc, Enargite. Enargite is one of the most 
difficultly soluble copper minerals; pure samples of tetrahedrite and 
tennantite are difficult to obtain and there are not many data on the 
solubility of these two minerals. Apparently tennantite is relatively 
insoluble in acid ferric sulfate, and tetrahedrite is somewhat more 
soluble. 

Ammonia. Native copper and copper carbonates (malachite and 
azurite) will dissolve in ammoniacal solutions to give the familiar 
deep blue solution in which the copper is present as a complex copper- 

6 Sullivan, J. D., op. cit., p. 524. 



310 HYDROMETALLURGY 

ammonium ion. The solvent used is ammonium carbonate formed by 
the dissolution of NH 3 and C0 2 gases in water, 

2NH 3 + CO 2 + H 2 ^ (NH 4 ) 2 C0 3 ^ 2NH 4 + + C0 3 ~~ 

and there is an excess of ammonia dissolved in the solution which 
hydrolyzes to form ammonium hydroxide. Thus the two effective 
reagents in the solution are ammonium carbonate (NH 4 ) 2 C0 3 and 
ammonium hydroxide NH 4 OH. 

The natural carbonates azurite and malachite contain copper in 
the cupric state, and these minerals dissolve directly in the solution 
to form the soluble cupric ammonium carbonate. 

Azurite: 

Cu 3 (OH) 2 (C0 3 ) 2 + (NH 4 ) 2 C0 3 + 10NH 4 OH 

-*3Cu(NH 3 ) 4 CO 3 + 12H 2 O 

Malachite: 
Cu 2 (OH) 2 C0 3 + (NH 4 ) 2 C0 3 + 6NH 4 OH - 2Cu(NH 3 ) 4 C0 3 + 8H 2 O 

Native copper is first dissolved by cupric ammonium carbonate in 
the solution to form cuprous ammonium carbonate; this is a common 
type of reaction in which cupric salts oxidize metallic copper to form 
cuprous salts. 

Cu + Cu(NH 3 ) 4 CO 3 -* Cu 2 (NH 3 ) 4 C0 3 

After this step the cuprous compound is oxidized to the cupric am- 
monium carbonate by reaction with oxygen and ammonium carbonate. 

2Cu 2 (NH 3 ) 4 C0 3 + 4(NH 4 ) 2 C0 3 + O 2 

-4Cu(NH 3 ) 4 C0 3 + 4H 2 + 2C0 2 

The oxygen is absorbed from the air by passing the solution through 
aerating towers. Thus the effective reagents in the dissolution of 
carbonates are ammonium carbonate plus ammonium hydroxide. 
Carbonate ions are contributed to the solution from the minerals. In 
dissolving native copper the effective reagents are ammonium car- 
bonate and oxygen, with cuprous ammonium carbonate as an inter- 
mediate product, and no carbonate ions are supplied to the solution. 
Solutions must be aerated in dissolving native copper, but this is not 
necessary for dissolving copper carbonates. 

The complex ions Cu(NH 3 ) 4 ++ and Cu 2 (NH 3 ) 4 ++ may not always 
hold four ammonia groups for each copper ion in these ammonia- 
carbonate solutions. This does not affect the general nature of the 



PREPARATION OF ORE FOR LEACHING 311 

reactions; for example we may rewrite the last two reactions thus: 
Cu + Cu(NH 3 )nC0 3 -> Cu 2 (NH 3 ) n C0 3 

2Cu 2 (NH 3 ) n CO 3 + n(NH 4 ) 2 CO 3 + O 2 

-4Cu(NH 8 ) n C0 8 + nH 2 + (n - 2)C0 2 

The C0 2 formed according to these reactions would immediately 
react with any excess ammonium hydroxide to form ammonium 
carbonate. 

Ammonia leaching is generally less satisfactory than acid leaching 
and is practiced only on ores which cannot be leached with acids. 
These include (1) ores having a carbonate gangue, and (2) tailings 
or ores of native copper; native copper does not dissolve readily in 
acid or acid ferric sulfate, so the oxygen-ammonium carbonate solution 
is the only practical solvent. There have been two ammonia leaching 
plants on the American Continents Kennecott in Alaska and Calumet 
and Hecla in Michigan and only one elsewhere Bwana M'Kubwa 
in Northern Rhodesia. The Calumet and Hecla plant treats reclaimed 
tailings containing native copper, and the other two plants leach copper 
carbonate ores. The Kennecott plant has recently been closed down. 

Ammonia leaching has one particular advantage in that there is 
little or no fouling of the leach solutions by dissolved impurities. 
None of the common elements form ammonium complexes such as 
those formed by copper, so there is little chance of finding anything in 
the ores except the copper minerals which will be attacked by these 
ammoniacal solutions. 

Other Solvents. 6 Other solvents which have been employed for dis- 
solution of copper minerals include hydrochloric acid, sulfurous acid, 
ferrous chloride, and ferric chloride. Some of these have been 
employed in connection with various types of roasting, and some were 
designed to dissolve precious metals as well as copper. The solvents 
that we have described previously, however, are the only ones that 
are commercially important at the present time. 

PREPARATION OF ORE FOR LEACHING 

Crushing. Before the ore can be acted upon by the lixiviant (leach- 
ing solution) it must be crushed fine enough so that the liquid has 
access to all particles of the copper minerals. For low-grade oxidized 
ores the ore is crushed in stages to about a %-inch maximum size and is 
leached directly. These ores are porous enough M that the solvent 
penetrates the particles and gives satisfactory leaching results. 

6 Hofman, H. O., and Hayward, C R , op cit. 



312 HYDROMETALLURGY 

Grinding. Low-grade ores (containing about 1 per cent copper) 
cannot be economically treated by grinding because of the expense 
involved. Not only would there be the cost of the grinding itself, 
but the product would contain more " slimes " which would have to be 
leached by agitation instead of percolation. Higher-grade ore, such as 
that at Katanga, can be ground in ball or rod mills and then leached 
in agitators, but most of the leaching ores do not contain copper enough 
to warrant the additional expense. All the large acid leaching plants 
except Katanga prepare the ore by fine crushing (% inch) without 
subsequent grinding. 

Classification or Washing. When the crushed or ground ore con- 
tains too many slimes for direct percolation leaching it must be sep- 
arated, into sand and slime portions; the washed sand is leached in 
vats by percolation and the fine slimes are treated in agitators. 
Whether classification is necessary or not depends to a large extent 
upon the amount of " natural slimes " in the ore, as the crushing of 
hard coherent rock to % inch would not produce enough fines to inter- 
fere with percolation. 

Roasting. The purpose of roasting a copper ore previous to leaching 
would be to convert sulfides to water soluble sulfates and acid soluble 
oxides, but for low-grade copper ores this would be prohibitively 
expensive. Roasting of high-grade copper concentrate to prepare it 
for leaching has been tried on experimental and semi-commercial 
scales only. 

Weathering or the natural oxidation of sulfide minerals by the 
combined action of water, air, and iron salts in the ore is sometimes 
used as a method of preparation. The ore is crushed and piled in 
heaps in the open where it is exposed to the action of the atmosphere. 
The actual leaching takes place concurrently with the weathering, and 
the water draining from the heaps contains the dissolved copper 
sulfate. Alternate wetting and drying of the ore speeds up the 
oxidizing reactions. The water may be added to the heaps from time 
to time, or the process may depend simply upon natural rainfall. 

Weathering is a slow operation and requires that the ore be kept in 
process for a long time; it is not subject to control or regulation, and 
it is wasteful unless the heaps are carefully drained and constructed 
to prevent the solution from seeping into the ground. Its advantages 
are its cheapness and the fact that most of the copper is oxidized to 
the water-soluble sulfate. 

Occasionally broken ore remaining underground will oxidize so that 
much of the contained copper can be leached out by passing water 
through the old stopes and collecting it in the lower levels of the mine. 



LEACHING METHODS 313 

Mine fires sometimes occur in sulfide ores, and the water circulated 
through these " fire zones " contains dissolved copper sulfate J 

LEACHING METHODS 

There are three principal methods of leaching copper ores (1) 
leaching in heaps or in underground stopes, (2) leaching in vats by 
percolation, and (3) leaching by agitation. We shall discuss these 
briefly here, and the details will be further illustrated in the description 
of plants which follows: 

Heap Leaching. Heap leaching has been practiced at Rio Tinto, 
Snain, since 1752, at Bisbee, Arizona, and on mine dumps at other 
places. This method is applicable to low-grade oies which cannot be 
profitably treated by other methods ; the ore must also contain sufficient 
pyrite to oxidize readily. 

The heap is built up so that the solutions readily drain away, and the 
heaps are built to permit uniform seepage throughout. It is best to 
build the heaps on a layer of fine tailings which form an impervious 
layer and prevent the solution from seeping into the ground. The 
surface of the heap is covered with shallow basins or pockets to hold 
the leach liquors and promote even distribution. 

The chemical action is largely oxidation brought about by the action 
of air and ferric salts. The piles are provided with ventilating flues to 
permit circulation of the air, and the ferric salts are produced by the 
oxidation of the pyrite and other iron sulfides. Copper is dissolved as 
CuS0 4 and is recovered in the solutions which drain from the bottom 
of the pile. 

The leaching agent is water, which is run onto the surface of the 
piles at intervals, plus the natural rainfall. In some places the spent 
liquor from the precipitating tanks is run back on the piles; it is a 
more effective leaching agent than water because it contains some 
ferric sulfate. 

The operations consist of alternate oxidation and leaching cycles, and 
for a large heap it may take several years to complete the leaching. 
Considerable heat is developed by the oxidation reactions, and the ad- 
mission of air must be regulated to keep the heap from taking fire. 

Leaching in Place. The leaching of broken ore in underground 
mines is similar to heap leaching in many respects. This method is 
used to recover copper from regions containing ore of too low grade 
for mining. Usually the ground to be leached is shattered by caving 
of nearby ore so that the rock is seamed and creviced and hence 
exposed to the action of air and circulating water. The leaching and 
oxidation cycles proceed in about the same way as in heap leaching, and 



314 



HYDROMETALLURGY 



the copper sulfate solution is collected in the lower workings and the 
metal precipitated; the solution may be first pumped to the surface 
or the precipitation may take place underground. 

In many copper mines the natural mine waters contain dissolved 
copper even though no special provision is made for leaching some 
particular region. The mine waters from the Butte mines, for example, 
are treated to precipitate the dissolved copper. 




(Courtesy Inspiration Consolidated Copper Company) 

FIG. 1. Leaching Plant at Inspiration. 
Leaching tanks and bridges visible at right. 

Percolation. Most of the large commercial leaching plants treat 
the ore in leaching vats. These are large square or rectangular vats 
from 60 to 175 feet on a side and 16 to 20 feet deep. The vats are 
constructed of reinforced concrete; and because the acid solutions have 
a corrosive effect on the concrete, the vats must be lined with lead 
or mastic (asphalt and sand). These tanks all contain false bottoms 
or filter floors set above the tank bottom to protect the bottom lining 
from damage when charging or excavating ore and tailings. These 
false bottoms are usually made of 2-inch planks supported on timber 
uprights; these planks are drilled with a number of small holes to 
permit passage of the solutions. Tank bottoms contain openings 
through which liquor can be pumped into the tank or can be drained 
from the tank. 



PERCOLATION 315 

As far as the ore is concerned, the leaching operation is a batch 
process. New ore is charged into an empty tank by means of the 
loading bridge (Fig. 1), it is treated by a series of leaching and 
washing solutions, drained, and finally excavated and sent to the 
dump. This leached material is the tailing. Commercial leaching 
vats will hold from 5,000 to 10,000 tons of ore per charge. 

There is no " flow " of ore through the leaching plant it simply 
goes into one tank and stays there until it is completely leached; then 
it is removed and another charge added. The flow of leaching solutions, 
however, is generally much more complicated and a given batch of 
solution may pass through several tanks in one cvcle All these large 
leaching plants employ a sulfuric acid leach and electrolysis for pre- 
cipitation of the copper, and the leaching solutions and electrolyte are 
essentially the same. Spent electrolyte is used for leaching the ore, 
and the resulting solution passes to the tank houses where part of 
the copper is deposited and an equivalent amount of sulfuric acid is 
generated. The spent electrolyte (depleted in copper and enriched 
in acid) then returns to the leaching tanks. In other words, the 
liquid circulates in a closed system through the leaching plant and 
the electrolytic tanks. 

There are two principal operations to be performed in these leach- 
ing vats (1) the ore must be kept in contact with the solution long 
enough to dissolve the copper minerals, and (2) the copper-bearing 
solution must be separated from the tailing. If a tank containing 
9000 tons of dry ore was allowed to drain completely, the wet tailings 
would still contain about 1000 tons of liquid, and if this was " strong 
solution," i.e., high in copper, it might contain up to 30 per cent of 
the total copper dissolved from the ore. There are two kinds of 
tailing losses in leaching operations chemical loss, which represents 
the copper that is undissolved, and mechanical loss of copper which is 
dissolved but which remains m the liquid entrained in the wet tailing. 

Filtration is out of the question for treating these huge masses of 
low-grade material, and the entrained copper solutions must be washed 
out of the ore while it is in the tank. The reason for using a number 
of different leaches and washes on each vat of ore is simply to obtain 
both efficient dissolution and washing and thus reduce both chemical 
and mechanical losses of copper to the practical minimum. 

It is difficult to explain the flow of solutions in a leaching circuit 
without reference to a specific flowsheet, and we shall present descrip- 
tions of leaching plants in a later section, where it will be possible to 
consider these questions in more detail. 

The liquid may flow or " percolate " continuously through a given 



316 HYDROMETALLURGY 

tank during the leaching, or the liquid may be run on, allowed to soak 
for a period, and then drained off. When liquid enters a tank through 
the bottom and overflows at the top, the system is known as upward 
percolation; in downward percolation the liquid enters at the top and 
is withdrawn at the bottom. 

The tanks may be connected in series with the leaching solutions 
flowing from one tank to the next in a counter current system. The 
spent electrolyte, plus the new acid, first is placed on ore which has 
been almost completely leached; when this solution (strongest in acid) 
is removed the ore is completely leached and is ready for washing 
and excavation from the vat. The liquid from this vat passes to 
the next, which is less completely leached, and so on through the 
series. Newly charged ore is first leached by solution which is low in 
acid and high in copper. Thus the " flow " of ore and solvent (actually 
the ore does not " flow ") are in opposite directions; the acid solution 
acts on successively richer and richer ore, increasing its copper content 
and decreasing its acid content as it goes. From the last tank the 
copper-rich solution passes to the electrolytic plant. 

The leaching may be conducted as a batch process by means of which 
a series of leach solutions are placed on the ore in one tank, allowed 
to soak for a period, and then withdrawn. In this type of leaching 
the tanks are not connected in series. 

Whatever leaching method is used, the dissolved copper and residual 
acid must be washed from the tailings. This is done by a series of 
washes with weak solutions and finally with water. Some of the 
first washes contain copper enough to join the main electrolyte; the 
last washes are usually treated with scrap iron to cement out the 
copper and are then discarded. 

Agitation. When the ore to be leached is of such a nature that it 
will not permit free passage of solution through the interstices, it 
must be agitated and kept in suspension in the leaching tanks until 
the copper is dissolved. Agitation tanks are much smaller than the 
tanks or vats used m percolation leaching and must be equipped with a 
mechanical stirrer or air lift to keep the pulp in suspension. Mechan- 
ical agitators contain rotating paddles for agitation. Pachuca tanks 
are vertical cylindrical tanks containing a pipe which is coaxial with 
the tank; compressed air is introduced at the bottom of the pipe, and 
the rising column of air bubbles reduces the density of the column of 
pulp which is forced out the top of the pipe by the pressure of the 
denser pulp surrounding the pipe. The pulp drops back into the tank 
and is thus kept in circulation. 

The separation of solid and liquid in slime leaching is a different 



AGITATION 



317 



problem than the simple draining that suffices for sand leaching. The 
separation is usually carried out in thickeners (Fig. 2), which are 
essentially cylindrical settling tanks. Pulp flows continuously into the 



Clear Overflow at 
/Tank Periphery 




(Courtesy The Dorr Company, Inc.) 

FIG. 2. Section Through a Dorr Thickener. 

thickener, and the solids settle slowly to the bottom of the tank; slow- 
moving arms sweep the thickened pulp toward the center of the 
tank where thickened pulp or spigot product is withdrawn continuously. 




(Courtesy The Dorr Company, Inc.) 

FIG. 3. A 325-foot Dorr Traction Thickener. 

Clear liquid overflows at the top of the thickener. Thickeners take a 
dilute pulp and resolve it into two products a clear solution and a 
thickened pulp. 



318 HYDROMETALLURGY 

When thickeners are used for washing a pulp they are generally 
used in series in a counter current washing system. The wash water 
flows through the thickeners in a direction opposite to the flow of pulp. 
The thickened pulp from one thickener is agitated (repulped) with 
weak solution; thickened again, repulped with a weaker solution, and 
so on until the values remaining in the liquid in the spigot product are 
too low to warrant use of another thickener. The spigot product from 
the last thickener is wasted as tailing. Wash water is added to the 
last thickener in the series, the overflow from this thickener is repulped 
with a spigot product to make the feed for the next to the last 
thickener, and so on. Figure 16 shows an example of a countercurrent 
slime washing system. 

Filters may be used for more efficient removal of liquid from slimes, 
but this is not common in copper leaching. 

TREATMENT OF THE LEACH SOLUTION 

After the leaching and washing have been completed, the pregnant 
solution passes on to precipitation, but some pregnant solutions must 
be purified before the precipitation stage. The solution from copper 
leaching (even from slime leaching) is usually quite free of suspended 
matter, so that it is not necessary to clanfy the solutions. It may be 
necessary to pass it through a settling tank, but it need not be 
filtered. 

When copper is to be precipitated on scrap iron the solution requires 
no special treatment but goes directly to the precipitation plant. 
However, when the leach solutions are in closed circuit with electro- 
lytic cells it is necessary to remove impurities picked up in leaching 
before the solution passes to the electrolytic cells. The most harmful 
of these soluble impurities are iron, chlorine, nitrates, and molybdenum. 
Rather elaborate installations are often required for purification of the 
electrolyte, and we shall consider these in connection with electrolytic 
precipitation. 

If the ore contains no soluble sulfate minerals there will be a 
continual loss of acid in tailings, discarded wash water, cementation 
launders, etc., plus that consumed in dissolving gangue minerals. It 
will be necessary, then, to introduce fresh acid into the circuit. On the 
other hand, if the ore contains copper sulfate minerals, the leaching- 
electro lysis system will generate a certain amount of free acid; this 
may exceed the acid loss so that excess acid is produced. 

CHEMICAL PRECIPITATION 

Metallic copper may be precipitated from pregnant solutions by the 
addition of certain reagents and, if the solutions are ammoniacal, 



CEMENTATION 319 

copper oxide is precipitated by simply boiling the solution. Many 
different precipitants have been tried, but only two methods are of 
commercial importance (1) cementation of sulfate solutions on 
iron and (2) boiling ammoniacal solutions to precipitate copper oxide. 
Cementation. Cementation is a rather simple procedure which de- 
pends on the fact that a metal can be displaced from solution by a less 
noble element. Thus if a piece of metallic iron is placed in a solution 
of copper sulfate the iron dissolves and copper precipitates according 
to the reaction 

CuS0 4 + Fe -> FeS0 4 + Cu 

Iron is the only metal used in practice for copper precipitation. 
Commonly scrap iron is used for this purpose although sponge iron 
produced by the direct gaseous reduction of solid iron oxide is a more 
rapid and efficient precipitant. 

The scrap iron is loaded into tanks, towers, or launders, and the 
leach solution slowly flows or trickles over its surface. The deposit 
of cement copper forms as a loosely adherent granular deposit on the 
iron pieces. This is dislodged at intervals by shaking or washing and 
is flushed into settling tanks. When a sufficient supply of cement 
copper has accumulated it is excavated and shipped to a smelter 
for treatment. As a rule it is charged into the reverberatory furnace, 
but especially pure cements may be charged into the converters or 
anode furnaces. 

If the scrap is badly oxidized the cement copper will contain a 
large amount of this oxide and will be quite impure ; cement copper may 
contain from only 60 to over 90 per cent copper, the purity of the 
deposit being highest with clean scrap and clean solutions. 

Scrap from " tin " cans is usually detinned before using, and heavier 
scrap is burned to remove any grease which would prevent the 
solution from wetting the metal. Precipitation with sponge iron is 
conducted in agitation tanks, and this method can be used with 
certain types of scrap iron. 

According to the precipitation equation, one mol of copper is pre- 
cipitated for each mol of iron, which would mean that one pound of 
precipitated copper would require 5 % 4 or 0.875 pound of iron. Ac- 
tually the consumption is greater than this, and commonly 1 to 2 
pounds of iron is required to precipitate 1 pound of copper. Free 
acid and ferric salts in the solution will also consume iron. 

H 2 S0 4 + Fe -> FeS0 4 + H 2 
Fe 2 (SO 4 ) 3 + Fe -> 3FeS0 4 



320 HYDROMETALLURGY 

Cementation has a number of advantages which may be listed as 
follows : 

1. The process is relatively cheap and simple, especially ii an 
adequate supply of scrap iron is available. It requires little super- 
vision and no elaborate equipment. 

2. It can be used on copper sulfate solutions of any strength and 
particularly on solutions which contain too little copper to be 
electrolyzed successfully. 

3. It can be used to " strip " solutions of all but a trace of copper 
so that the solution can be discarded. Electrolytic methods cannot be 
successfully used to strip solutions of their copper content. 

Disadvantages of the cementation process are: 

1. It produces a finely divided precipitate of metallic copper which 
must be melted and refined ; often this cement copper is rather impure. 

2. The metallic iron will consume any free acid in the solution, and 
therefore it cannot be used on acid solutions without excessive con- 
sumption of iron. Moreover, the acid is destroyed so that this 
method cannot be used in closed circuit with an acid leach. 

Cementation is most commonly used (1) on leach solutions from 
heap leaching and underground leaching, and on mine waters, and 
(2) for stripping lean wash solutions from acid leaching. 

Other Chemical Methods. Many other processes have been de- 
veloped to precipitate copper either as a compound or in the metallic 
state. All these methods have one disadvantage in common with 
cementation they yield a precipitate which must be resmelted 
(whether it is metallic copper or a copper compound). Some of these 
precipitants, however, have an advantage in that they do not consume 
the free acid in the solution (H 2 S and S0 2 , for example). The prin- 
cipal use of some of these methods is in the purification cycle in con- 
nection with electrolytic precipitation. 

Hydrogen Sulfide. Hydrogen sulfide (H 2 S) will precipitate copper 
as the sulfide, but the precipitant is difficult to handle and requires 
further smelting. 

Sulfur Dioxide. Sulfur dioxide gas (S0 2 ) under pressure will pre- 
cipitate copper from solution, thus: 

CuS0 4 + SO 2 + 2H 2 - Cu + 2H 2 S0 4 

This method yields a precipitate of metallic copper and generates 
sulfuric acid. 

Lime. Burnt lime will precipitate copper as the hydroxide, but 
this gives a bulky precipitate which will be contaminated with other 
base metals and CaS0 4 . 



ELECTROLYTIC PRECIPITATION 321 

Precipitation from Ammoniacal Solutions. The copper dissolved 
in ammonia leaching exists in the form of a complex cupric ammonium 
carbonate which is soluble in excess ammonium carbonate and am- 
monium hydroxide. Some of the equilibria found in these solutions 
are indicated by the following reactions : 



Cu(NH 3 ) 4 C0 3 ^ Cu(NH 3 ) 4 ^+ + CO 3 " (1) 

Cu(NH 3 ) 4 C0 3 + 4H 2 ^ CuC0 3 + 4NH 4 OH (2) 

\\ 
4NH 3 + 4H 2 

(NH 4 ) 2 C0 3 + 3H 2 ^ 3NH 4 OH f H 2 C0 3 (3) 

jr \\ 

3NH 3 + 3H 2 H 2 O + CO 2 

H 2 O + CuCO 3 ^ CuO + H 2 C0 3 (4) 

jr 

H 2 O + C0 2 

When such a solution is boiled the NH 3 gas is expelled and this causes 
reactions 2 and 3 to proceed to completion in the direction of the upper 
arrows. Thus both ammonium carbonate and cupric ammonium 
carbonate are hydrolyzed to yield carbonic acid and a precipitate 
of cupric carbonate (or basic carbonate). Carbonic acid readily de- 
composes into C0 2 and H 2 if there is not enough ammonia in 
solution to fix it as ammonium carbonate, and the slightly soluble 
C0 2 gas escapes from solution. Further boiling hydrolyzes the copper 
carbonate and drives off the CO 2 , leaving a precipitate of CuO, as 
shown in equation 4. 

The net effect of boiling these solutions is to expel the ammonia and 
carbon dioxide and cause the copper to precipitate as a carbonate, 
which becomes black copper oxide on further boiling. The ammonia 
and C0 2 can be recovered from the gases to be used again as a leaching 
agent. 

The precipitated copper oxide is then shipped to a copper smelter, 
where it is charged to the smelting or refining furnace. 

ELECTROLYTIC PRECIPITATION 

The most widely practiced method for the precipitation of the 
copper dissolved in acid leaching is by electrolysis, using insoluble 
anodes. This method has one outstanding advantage over all other 
precipitation methods it yields directly a product (cathode copper) 
which is of the same quality as the cathodes produced by electrolytic 
refining. 



322 HYDROMETALLURGY 

In some respects the process resembles the electrolytic refining of 
copper the tanks and electrical connections are similar to the 
multiple refining process, the same type of starting sheets are used, 
and the finished cathodes are treated in the same way as cathodes 
from electrolytic refining. There are, however, several fundamental 
differences between the two systems, as we shall see. Briefly, these 
differences may be summarized as follows: 

1. Insoluble anodes are used, and there is no appreciable corrosion 
of them (a good anode may last more than 10 years), and hence no 
" anode mud " is formed. 

2. The copper in the electrolyte comes from the leaching plant, 
which is in closed circuit with the electrolytic cells, and because there 
is no copper dissolved from the anodes, the electrolyte becomes de- 
pleted in copper and its free acid content increases as it passes through 
the tank house. In the leaching cycle the opposite effect is found 
acid is used up and more copper is dissolved. 

3. Current efficiency is generally lower, voltage is much higher, and 
the power consumed per pound of cathode copper is much greater 
(Table 3) . The cathode current density is less than that used in 
refining. 

4. Concentrations of dissolved copper and free acid in the electrolyte 
are generally less than in refinery electrolytes, and the resistance of 
the electrolyte is greater. 

5. The impurities found in the electrolyte are different from those 
found in refinery electrolytes, and the problem of purifying the 
solution is altogether different. 

6. Electrolytic tanks are generally longer than multiple refining 
tanks, but anodes and cathodes have about the same dimensions. 

The Cell Reaction. As we have mentioned before, electrolysis con- 
sists of two equivalent and opposite chemical reactions oxidation at 
the anode and reduction at the cathode. In refining of copper the 
two reactions practically balance one another. Copper is plated on the 
cathode and an equivalent amount of copper is dissolved (corroded, 
oxidized) at the anode. With insoluble anodes, however, the cathode 
reaction is the same as in refining: 

Cu++ + 2(e) - Cu 
but the anodic reaction is different: 

S0 4 ~" + H 2 - 2(e) - H 2 S0 4 + f 2 
The net cell reaction is the sum of these two, or 

CuSO 4 + H 2 O - Cu + H 2 SO 4 + 2 \ 



THE CELL REACTION 323 

The electrolysis liberates metallic copper at the cathode and gaseous 
oxygen at the anode; for every mol of CuS0 4 decomposed a mol of 
free H 2 S0 4 is formed in the electrolyte. 

The cell reaction if endothermic, i.e., if it proceeds from left to right, 
energy is absorbed: 

CuS0 4 + H 2 -> Cu + H 2 S0 4 + ^0 2 - 55,030 cal 

In other words, copper will dissolve in oxygenated sulfuric acid and 
liberate 55,030 calories of heat for every gram-mol of copper dissolved. 
For the reverse reaction, as it takes place in electrolysis, 55,030 
calories of heat are absorbed for every rnol of copper that is deposited, 
and this energy or its equivalent must be supplied by some external 
source. In electrodeposition this is supplied as electrical energy, and 
if we assume that electrical energy is converted directly into " chemical 
energy " we can derive the following relations. 

1 gram-calorie = 4.186 joules or watt-seconds 

Hence 55,030 calories represents 55,030 X 4 186 =- 230,000 watt-seconds 
of electrical energy that must be supplied for each gram-mol of copper 
liberated. From Faraday's law we know that the deposition of 1 mol 
of divalent copper (2 equivalents) requires 2 faradays or 2(96,500) 
coulombs of electricity. An expression of the amount of electrical 
energy is the product of two factors a capacity factor (measured in 
coulombs) and an intensity factor (measured in volts) ; 1 joule is a volt- 
coulomb, or the amount of energy represented by 1 coulomb falling 
through a potential of 1 volt. 

The voltage in our example is unknown ; let us set it equal to V, and 
then we may write: 

2(96,500) V = 230,000 

Whence V = 1.195 volts. 

This voltage is the decomposition potential or cherttical potential of 
the indicated reaction, and it means that the voltage drop across the 
cell must be at least 1.195 volts if the reaction is to take place as 
indicated. If an insoluble anode is used, and there are no other re- 
actions possible but the one indicated, then no current will flow through 
the cell unless the impressed voltage from anode to cathode is greater 
than 1.195 volts. After the voltage exceeds .his value the current 
will flow, and the current flowing will follow Ohm's law i.e., the 
current flowing will be directly proportional to the voltage above 
volts. 



324 



HYDROMETALLURGY 



This method of calculation is known as Thomson's rule and is based 
on the assumption that the electrical energy and heat energy involved 
in any given reaction are equal. We may write Thomson's rule as 



E 



U 

njF 



where E is the decomposition voltage, n is the valence, U is the heat 
of the reaction in calories, F is 96,500 coulombs, and ;' is the factor 

for converting joules to calories = 0.239 = 

4.186 

Thomson's rule is important because it illustrates the meaning of 
decomposition potential in terms of the energy change involved in a 
given reaction. It is not strictly accurate, however, because in a 
given reaction the chemical energy may not all be converted into 
electrical energy or vice versa; some heat energy may be absorbed 
from or given up to the surroundings. The correct formula is the 
Gibbs-Helmholtz equation: 

E ** njF + T dT 

where T is the absolute temperature of the cell. This equation is 
essentially Thomson's rule plus a correction factor which takes into 



First Class Conductor 



b. Second Class Conductor 
(Insoluble Anode) 



FIG. 4. Current-Voltage Curves for First-Class and Second-Class Conductors. 

The curve for an electrolytic cell uamg a soluble anode (refining) would resemble a, i.e., there would be 

no decomposition potential. 

account the heat evolved or absorbed; the temperature coefficient 
dE/dT must be known if the Gibbs-Helmholtz equation is to be used. 
This may be either positive or negative. 

The decomposition potential may be determined experimentally by 
gradually increasing the voltage and plotting current against voltage 
(Fig. 4). The first voltage at which an appreciable current begins 
to flow is the decomposition voltage. 



ANODES 325 

The decomposition potential for copper sulfate between insoluble 
electrodes (platinum) as determined experimentally is given as 1.49 
volts, 7 for a normal solution of copper sulfate. 

Cell Voltage. The voltage across the deposition cell must be about 
1.49 volts plus sufficient additional voltage to give the required current 
density, assuming that the concentration of 'he copper sulfate is 
normal and that the anode has the same characteristics as a platinum 
anode. Experimental determinations of decomposition potentials must 
also include the gas and metal overvoltages. 

The overvoltage is the additional voltage required to deposit a metal 
or liberate a gas at an electrode; it depends up( n the metal or gas to 
be liberated and also upon the nature of the electrode. In general, the 
metal overvoltages are rather small, but gas overvoltages are larger. 
In copper electrolysis we are interested in the anode overvoltage of oxy- 
gen; this will vary with the nature of the electrolyte, the composition of 
the anode, and the current density. At a current density of 1.0 ampere 
per square decimeter the oxygen overvoltage will range from 0.5 to 1.0 
volt, 8 depending on the material in the anode. 

The total cell voltage will then consist of three parts: (1) the decom- 
position potential as calculated from the Gibbs-Helmholtz equation, 
(2) the oxygen overvoltage at the anode, and (3) the voltage required 
to overcome the contact resistances and the ohmic resistance of the 
electrolyte. These will be approximately (1) 1.20 volts, (2) 0.30 to 
0.60 volt, and (3) 0.20 to 0.30 volt, respectively, or a total of 1.70 to 
2.10 volts. 

Voltages used in practice range from 1.86 to 2 24 volts. The voltage 
will depend upon the nature of the anode, and it will vary slightly from 
time to time, depending on the condition of the anode surface. 

Anodes. The function of the anode is simply to serve as a con- 
ductor to bring the current into the electrolyte; it should not react with 
the acid in the cell nor be oxidized by the liberated oxygen; and its 
oxygen overvoltage should be as low as possible. The anodes should 
also have sufficient mechanical strength and not be too brittle. 

A number of different materials have been employed as anodes 
antimonial lead, copper silicide and other copper alloys, magnetite, and. 
cast irons. At present the most common material is antimonial lead,.* 
but part of the anodes at the Chuquicamata plant are of " Chilex," an , 
alloy of copper, silicon, iron, and lead, with small amounts of tin and 
other metals. The anodes may be either cast O A - rolled, and they may 

7 Mantell, C. L., Industrial Electrochemistry, 2d ed., p. 61, McGraw-Hill Book 
Co., New York, 1940. 

8 Creighton, H. J., and Koehler, W. A., op. cit., Vol. 2, p. 47. 



326 



HYDROMETALLURGY 



be supported by cast lugs (like those on copper refinery anodes), cop- 
per lugs, copper inserts, or riveted copper bars. The anodes may be 
solid slabs, or they may be in the form of a grid to save weight and 
material. As we have noted, they are about the same size as refinery 
anodes. 

Cathodes. The cathode deposit is formed on starting sheets of the 
same type as those used in refining. The starting sheets are made in 
stripper cells which usually employ soluble copper anodes and operate 
in much the same way as stripper cells in refineries; insoluble anodes 
are used in some stripper cells. 

Electrolyte. Table 1 gives the typical analysis of an electrolyte, 
and Table 2 shows the composition of electrolyte before and after it 
passes through the deposition cells. 

TABLE l a 
ANALYSIS, IN GRAMS PER LITER, OF SPENT ELECTROLYTE AT CHUQUICAMATA 

Cu 

Acid 

Cl 

Fe 

SO 4 

HN0 3 

a Mantell, C L , op cit., pp 292, 301 

TABLE 2 a 

ANALYSES, IN GRAMS PER LITER, OF ELECTROLYTES AT 
CHUQUICAMATA AND INSPIRATION 



14.44 


Al 


1 75 


K 


2.30 


78.54 


As 


242 


Pb 


Oil 


21 


Sb 


0.009 


Sn 


004 


4 60 


Ca 


475 


Mn 


135 


144 83 


Mg 


147 


Mo 


0.248 


3 12 


Na 


6 74 









Chuquicamata 


Inspiration 


Entering 
Tank House 


Leaving 
Tank House 


Entering 
Tank House 


Leaving 
Tank House 


Cu 
Free H 2 SO 4 

cu 

Total Fe 
Ferric Fe 
Ferrous Fe 


22 
55 
0.10 
2 5 
0.9 
1.6 


18 
61 
0.10 
2.5 
1.3 
1.2 


32.6 
22.2 

16 8 
4 2 


28.1 
26.2 

8.4 



a Mantell, C L., op cit , p 294 



Comparing these tables with the data in Table 3, Chapter VIII, 
which gives the composition of refinery electrolytes, we see that the 
^electrolytes used in electrowinning contain about half as much dissolved 
copper and only one-third to one-fifth as much free acid. These have 



ELECTROLYTE 327 

a lower conductivity than refinery electrolytes, and they are also less 
corrosive. 

These electrolytes are rather complex (Table 1) because of the 
number of substances dissolved from the ore; the concentration of these 
will build up in the closed circuit, so provision must be made to keep 
them below certain critical concentrations. The most harmful im- 
purities found in these solutions are as follows : 

Ferric Iron. Ferric iron oxidizes copper at the cathode according to 
the reaction 

2Fe+++ + Cu -* 2Fe++ + Cu"^ 
or 

Fe 2 (S0 4 ) 3 + Cu - 2FeS0 4 + CuS0 4 

This corrosive effect is very pronounced and diminishes the current 
and power efficiencies. Addicks 9 found that in a solution containing 
0.75 per cent ferric iron (about 9 grams per liter) at 51 C no cathode 
deposit was formed. 

Under certain conditions the ferrous ions may be oxidized back to 
ferric ions in the electrolytic cells; this causes an increase in cathodic 
corrosion by the resultant ferric ions. The oxidation may take place 
at the anode: 



Fe++ - (e) 

but as this requires a higher potential than the normal anode reaction 
S0 4 ~ + 2H 2 - 2(a) -> H 2 S0 4 + i0 2 

with moderate Fe+ + concentration and well-circulated electrolyte there 
may be no anodic oxidation of the ferrous iron. 

Ferrous iron is generally considered harmless and no effort is made 
to remove it, but all iron is simply reduced to the ferrous state. If 
the ore to be leached contains sulfides, the presence of ferric iron aids 
in the dissolution of copper, as we have noted. The practice at any 
particular plant, therefore, will depend upon whether or not it is de- 
sired to dissolve sulfide copper; ferric iron will improve the copper dis- 
solution, but the tank house efficiency will suffer correspondingly. 

Nitrates in the solution, particularly when catalyzed by small 
amounts of molybdenum, may cause a vigorous oxidation of ferrous 
ions. 

Ferric iron in solution may be reduced to ferrous iron in one of two 
ways: 

* Addicks, L., in Creighton, H. J., and Koehler, W. A., op. cit., Vol. 2, p. 198. 



328 HYDROMETALLURGY 

1. Bringing the solution in contact with S0 2 gas in reaction towers. 
The reaction involved is 

Fe 2 (S0 4 ) 3 + S0 2 + 2H 2 -> 2FeS0 4 + 2H 2 S0 4 

This reaction reduces the ferric iron below 0.4 per cent and generates 
sulfuric acid. 

2. Passing the solution over cement copper. This method can be 
used to remove all but a trace of ferric iron, but no acid is generated. 
This method is generally used in connection with dechloridizing op- 
erations. The reaction involved is 

Cu + Fe 2 (S0 4 ) 3 -> 2FeSO 4 + CuS0 4 

Chlorine. Chlorine, especially when present in amounts over 0.5 
gram per liter, causes difficulties in tank house operation. The anodes 
are attacked and corroded, insoluble cuprous chloride deposits on the 
cathode, and chlorine gas is liberated into the tank house atmosphere. 
Solutions are dechloridized by passing them over cement copper, which 
precipitates the chlorine as insoluble cuprous chloride (chlorine exists 
in solution as cupric chloride) 

Cu + CuCl 2 -> 2CuCl 

This reaction goes practically to completion in a few minutes if finely 
divided copper is present in excess. Cuprous chloride is practically 
insoluble in these dilute solutions, but it is soluble in a hot strong brine 
of ferrous chloride. The copper precipitated as CuCl is recovered by 
dissolving in strong ferrous chloride and then reprecipitating it as 
cement copper by cementing on iron. Note that the operation of this 
process results in the formation of excess cement copper, because in the 
cuprous chloride there is twice as much copper as the copper required 
to precipitate it. 

Nitrates. Under certain conditions the nitrate ions cause severe 
oxidizing conditions in the electrolyte; ferrous ions are oxidized; anodes, 
cathodes, and lead pipes are oxidized; and fumes of nitrous oxide are 
evolved from the tanks. Molybdenum serves to catalyze these reac- 
tions, and when they get a start the nitrous oxide acts as a self- 
catalyzer. Nitrate and molybdenum are controlled by stripping and 
discarding a portion of the electrolyte; charging the electrolyte with 
SO 2 gas (a reducing agent) minimizes the oxidizing action of the 
nitrates. 

Other Impurities. Most of the other impurities in these solutions 
are relatively harmless, and the discarding of a certain amount of 



ELECTROLYTE 



329 



solution at regular intervals keeps their concentration within safe 
limits. 

Purification with Limerock. 1 At the leaching plant of the Andes 
Copper Mining Company, Potrerillos, Chile, the solution from the 
leaching tanks is agitated with limerock and basic cop >er carbonate, 
which is obtained by stripping (precipitating) wane solutions with 
limerock. This precipitates most of the ferric iron, arsenic, phos- 
phorus, molybdenum, and about 20 per cent of the aluminum; at the 
same time the copper in the carbonate precipitate is redissolved. This 
purification completely neutralizes all the free acid remaining in the 
leach solution. Practically all of the limerock enters the system at 




(Courtesy Inspiration Consolidated Copper Company) 

FIG. 5. Interior of Tank House at Inspiration. 

the point where the waste solutions are stripped, and only a small 
excess is needed for the purification. 

The neutralized solution is filtered and sent through the dechloridiz- 
ing plant. Fresh acid is then added to bring the free acid content up 
to 10 or 15 grams per liter, and the solution proceeds to the tank house. 

General. The electrolyte is maintained at a lower temperature (30 

10 Callaway, L. A., and Koepel, F. N., Metallurgical Plant of the Andes Copper 
Mining Company: Am. Inst. Min. & Met. Eng. Trans., Vol. 106, p. 709, 1933. 



330 



HYDROMETALLURGY 



to 40 C) than is used in refining practice; higher temperatures would 
decrease the resistance of the electrolyte and improve the leaching 
action, but they would also increase the corrosive and oxidizing action 
of the electrolyte. 

The circulation of the electrolyte is much more rapid than is com- 
mon in refining; this is permissible because there is no anode mud to be 
stirred up. The circulation rate will range from 25 to 200 gallons per 




(Courtesy Inspiration Consolidated Copper Company) 

FIG. 6. Starting Sheet Section, Inspiration Tank House. 

minute, depending on the arrangement of the circulation cascades. A 
film of oil is usually used to cover the electrolyte to prevent the spray- 
ing of acid into the atmosphere by the escaping bubbles of oxygen. 

Note (Table 2) that only a small part of the dissolved copper is re- 
moved from the electrolyte in the tank house. If it is desired to main- 
tain the purity of the deposit and efficiency of deposition, the copper 
content must not fall below a certain minimum figure. Electrolysis is 
not suitable for completely stripping the copper from solution; cementa- 
tion is commonly used for this purpose. 

Electrolytic precipitation is best suited for treatment of the relatively 
high copper, high acid solutions obtained by acid leaching. The low- 
copper solutions containing little or no free acid, such as are obtained 



LEACHING-CONCENTRATION 



331 



Current. The current densities used in electrowinning are lower than 
those used in refining from 5 to 13 amperes per square foot of 
cathode surface. The current efficiency is also lower; it will be about 
85 to 90 per cent when the ferric iron is kept dowr but only about 70 
per cent when ferric sulfate is used as a leaching a^ent for sulfides. 

Tanks are connected in about the same way as n multiple refining 
electrodes in parallel, tanks in series and the 'Valker system of 
connections is most common. 

Power Consumption. Because of the high voltage and lower cur- 
rent efficiencies, much more power is icquired th n m refining. Table 
3 (analogous to Table 1, Chapter VIII) shows me power consumption 
for different rell voltages calculated for 90 and 70 per cent current 
efficiencies. Comparison of the two tables shows that from 8 to 16 
times as much power is required per pound of copper m electrowin- 
ning. In extreme cases the electrowinning process might require 19 
times as much power as electrorefining 

TABLE 3 

ELECTRICAL REQUIREMENTS FOR COPPER DEPOSITION 
USING INSOLUBLE ANODES 





Cell Voltage 




Cunent Efficiency = 90% 




1 8 


1 9 


2 


2 1 


2 2 


Ampere-hours per pound of copper 


425 


425 


425 


425 


425 


Kilowatt-hours per pound of 












copper 


705 


807 


850 


892 


935 


Pounds of copper per ampere-day 


0565 


0565 


0565 


0565 


0565 


Pounds of copper per kilowatt-day 


31 4 


29 7 


28 2 


26 8 


25 6 




Current Efficiency = 70% 


Ampere-hours per pound of copper 


545 


545 


545 


545 


545 


Kilowatt-hours per pound of 












copper 


080 


1 035 


1 090 


1 143 


1 20 


Pounds of copper per ampere-day 


0440 


0440 


0440 


0440 


0440 


Pounds of copper per kilowatt-day 


24.4 


23 2 


22.0 


21.0 


20.0 



LEACHING-CONCENTRATION 

Resembling in principle the Katanga process of first reducing cop- 
per oxides to metallic copper and then removing the metallic copper 
by concentrating methods, is an acid-leaching process in which the 
" normal " order of clarification and precipitation is reversed. This 



332 HYDROMETALLURGY 

process involves (1) leaching the ore to put the copper into solution, 
(2) precipitating the copper by treating the pulp with iron, and (3) 
removing the precipitated copper from the pulp by flotation. 

This method is applicable to ores and old tailings which are not 
adapted to " standard " leaching methods for one or more of the fol- 
lowing reasons: 

1. The ore is of too low grade, and the resulting leach solutions too 
poor in copper to utilize acid leaching followed by electrolytic precipi- 
tation. 

2. The ore may contain a mixture of sulfides and oxides too much 
oxide for flotation alone, and too much sulfide for simple leaching. 

3. The material may not settle readily enough to make thickening 
and clarification practical this is particularly true of old weathered 
tailing dumps. 

We shall give a brief description of the plant of the Ohio Copper 
Company at Lark, Utah, which treats an old tailing containing 0.42 
per cent copper of which 22 to 25 per cent is water-and-acid-soluble. 
The oxide copper is dissolved, the copper precipitated, and the sulfides 
and metallic copper removed as a bulk flotation concentrate. A similar 
process is employed by the Miami Copper Company except that there 
sulfides and metallic copper are recovered separately in two flotation 
circuits. 

Ohio Copper Company. 11 Figure 7 shows a flowsheet of the opera- 
tion. Monitors wash the tailing from the dump, and the pulp flows 
through a surge tank into a washing trommel equipped with a 16-mesh 
stainless steel screen. Rock particles and debris remaining on the 
screen are rejected, and the undersize is pumped to the first of three 
conditioners connected in series. About 6 pounds of 60 Baume sul- 
furic acid per ton is used; this is all added to the first conditioner. Of 
the total copper in the tailing, about 125 per cent is water-soluble, and 
12.5 per cent acid-soluble. Thus about 12.5 per cent of the copper is 
dissolved before the pulp enters the conditioners, and 25 per cent is 
dissolved when the pulp leaves the last conditioner. 

The conditioners are wooden, rubber-lined tanks, and the ship-type 
propellers are rubber covered; the acid agitation serves four purposes: 

1. It dissolves the acid-soluble copper. 

2. The acid cleans the surfaces of sulfide particles and renders them 
more floatable. 

3. The tanks serve as conditioners for the flotation circuit; 0.025 
pound of " Minerec " (collector for sulfides) is added here. 

u Milliken, F. R., and Goodwin, Robert, Ohio Copper Company Tailings Re- 
Treatment Plant: Am. Inst. Min. & Met. Eng. Tech. Paper 1221 (Mining Tech- 
nology), July 1940. 



OHIO COPPER COMPANY 



333 




Legend 

/'Monitors 

B Frame Hydroseal 

4' x 4' Surge Tank 

Automatic Sampler 
48"x84"x 16 Mesh Trommel 
B Frame Hydroseal 
10'x 10' Conditioner 

8'x8'Cond!tioner 
e'xB'Precipitator 
32"x 2 Mesh Trommel 
(fi) Pulp Distributor 
@ 8 Cell No. 24 Denver Rougher 

8 Cell No. 24 Denver Cleaner 
@ 2"Sand Pump 
20' Thickener 
@ l"Diaphram Pump 
6 Disc 4' Filter 



Concentrate Tailings 

(Milliktn and Goodvrin, Am. Inat. Mm. and Met. Eng. Tech. Paper 1 881, M\n. Technology, July 1940) 

FIG. 7. Flowsheet of the Tailings Treatment Plant, Ohio Copper Company. 



334 HYDROMETALLURGY 

4. The conditioners act as surge tanks to prevent sudden changes 
in the density of the pulp flowing into the flotation circuit. 

From the acid agitation the pulp flows through three precipitating 
tanks in series. These are rubber-lined wooden tanks equipped with 
white cast iron ship-type impellers rotating at 250 revolutions per 
minute. The iron used is known as " Premt " or " shredded iron " 
and consists of sheet iron obtained from discarded tin cans shredded 
into pieces about 2 inches square. Under the impeller action this iron 
moves as a boiling mass near the bottom of the tank and gives good 
contact with the solution. The action is less vigorous near the top so 
that very little of the iron is carried over with the pulp. Iron is added 
by hand as needed. About 2 l / 2 pounds of iron is consumed per pound of 
copper precipitated this is higher than the consumption in the usual 
precipitation plant because the iron also neutralizes the free acid left 
in solution. With normal pulp flow the solution receives about 5 min- 
utes contact with the iron, which suffices to precipitate 82.9 per cent of 
the dissolved copper. 

The pulp passes over a trommel to remove the small amount of iron 
carried out of the last agitator and then to the flotation circuit. Pen- 
tasol xanthate (0.04 pound per ton) and amyl alcohol (0.15 pound per 
ton) are added as collector and frother respectively. Metallurgical 
results obtained are tabulated in Table 4. This process recovers 72 38 
per cent of the total copper in the original tailing in a concentrate as- 
saying 25.37 per cent copper. 

EXAMPLES OF PRACTICE 

To illustrate leaching methods we shall give brief descriptions of five 
commercial applications. These will include one operation leaching 
ore in place (Ray Mines), one ammonia leaching plant (Kennecott), 
and three acid-leach electrowmning plants (Chuquicarnata, Inspira- 
tion, and Katanga) . 

Ray Mines. 12 Leaching of mined-out areas at the Arizona property 
of the Ray Mines Division, Kennecott Copper Corporation, was started 
on January 20, 1937, and by July 1, 1938, 10,000,000 pounds of copper 
had been produced by this method. The leaching operations are con- 
fined to the western part of the ore body and extend over an area of 
about 10 acres. The ore mined under this area averaged slightly more 
than 1 per cent copper and was extremely high in pyrite. Above the 
ore was an unaltered zone of primary or protore averaging 125 feet 

12 Thomas, R. W., Leaching Copper from Worked-Out Areas of the Ray Mines, 
Arizona: Mining and Metallurgy, Vol. 19, No. 383, p. 481, 1938. 



OHIO COPPER COMPANY 



335 







1 


1 




8S 


8 




8 1 OS O OS 
-^ 00 00 00 




00 ^ 

co 10 


$ 




ss^ 


s 


5 









o 




CN t^ 

l^ (N 


8 




8 <N o;o;o; 




CN "O 

t CN 







^^ tN 


CN 








l 


d 
























a * 
-3 f-t 






p s 


OQ 
























a 


o> 






gpQ 


F2 




^< 00 


8 




8SJ5S^ 




o> 


s 




O OS CO 

o oo os 





i IM 


1 " H 




fc 






H 00 


8 




8 CO CO CO CO 
os os os 




^ 05 


$ 




CN CN CO 

OS 


OS 


5 S ** 


ti 


m 


^ 


Q 






?H 




1-( 














os -73 * 

-M w Q) 


w 


^5 




























^ g 


1 


+ 


3 






88 


8 




mil 




88 







888 


8 


T-J S ng 


o 




3 









CO CN 


3 

OS 




O) Os OS OS OS 
OS OS OS OS 




CO CN 


o> 

05 




CO CO OS 


S 


































* & *! 


8 


CO 


3 


























S 2 2 




o 


S 

-27 


S 




CO 




"*< 




O O 00 00 00 




CO i 


. 




*O i O 
SO CN 
CN 


S 


1|| 




p 




| 

a> d 







iO O 


O 




o o o o o 




3 







O 


o 


J.S.J 






<2 


























& 5 -| 


g 




O 3 


























fc i & 






j * 


























Cj &< >H Q 


fa 




15 


3 

3 




s 


Jo 




2 




o 


CO 




.0 


s 


* X 

"O 


g 




g-3 




1 




CN 


2 


55 


S ^c^ M 




2 


CO 




CO "O 


2 


a 

OJ 

'S 


w 














O 


















s 




PH 










H 
















S 


H 














<5 
































^ 


















W 




3 




J 


11 


8 


S 


8888 




1 


1 




111 


1 


5 


Tt< H 

n W 

S rt 




o 




w 


CO <M 


i 


O 
H 

C^ 


i O> i O *t 

o ce> ~ oo 

00 I s - t>- 00 





OS CO 
CO CS 


"* 

s 


CO 

p 


2 


o 

a 


w 

a 


g 1 


a 


"i 


3 
J 


O 


, 


CO 


Q 
K 
< 


^ M ^ (N O 

CO CN CO CO ^ 


i 


^ s 

CO 


s 


a 

H 


00 O <N 

O >0 CN 


1 


s 

03 
if 


^ ni 










jn 


ift O 





O 


o o o o o 


fe 


iO O 


o 




O O O 


o 





<< 














^ 


















CH 






























g. 
















X 


















S 











11 


s 


O 

< 

H 


*O O 1 O i i O 

o <o o o f"* 




II 


8 




5SS 


1 


1 1 






H 






-i CO 


2 


^ 


2 -* o o o 




~ 


o 




o w ^ 


CO 


J 


O 






























o 1 S 


o 






























28 


B 




J2 



























* a^ 


o 




" 












... 




. 






^. ^ 


. 


ev> ** 


^ 




^ 






8 







W MOC^ 




CO 


55 




N 


$ 


S 1& * ^ 

C -^ O M 


< 






























G * b- jf 


s 






























** S. ^ S 


i 


C 
O 


feS 














^ 







CN 




28 


o 


fe V fl Hfl 

| B 2 fl 


CG 


jj 


g 






"^ 


o 




lO ^ J* I-H 




^ 


11 




* * ^ 


*"* 


oj 2 


S 


<8 


^h 






O 







000 




o 


o 




O 


o 


s fl 'ft 


| 




S 






| 


8 




8 III 




1 


1 







1 




1 




fi 






S 


5 




< CO CO CO 




i 


CO 




1 s 





. "o a 


^ 






























S 2 h 8 


H 






























?h 


W 
^4 
















^ 










S 




"S g i-3 


S 
















g o j) 










a 




3 o C ti 






1 






Concentrate 
Mill tails 


j 




C CP *** T3 

* * G v 

i||!i 




Concentrate 
Flotation tails 


Flotation it^d 




Rougher tails* 
Cleaner tails* 
Trommel oven 


Mill tails 


a Trommel 
82.93 per 
Indicated 
' Some adc 



336 HYDROMETALLURGY 

in thickness and containing about 0.6 per cent copper, and above this 
was a 50-foot leached zone or capping. 

By 1933 there was no further mining in this section, and there was 
evidence that the remaining ore was broken and oxidized and that oxi- 
dation had extended into the protore zone. During 1935 and 1936 
there was considerable rainfall, and the copper content of the surface 
water after percolating through this broken ground averaged 1 per 
cent copper, or 83 3 pounds per 1000 gallons. As no further mining 
was contemplated in this region it was decided to prepare the section 
for leaching; an estimate of the copper content of the abandoned ore 
and the protore above it indicated that the ground amenable to leach- 
ing contained over 50,000,000 pounds of copper. 

Considerable underground work was required to prepare the area for 
leaching. Drainage drifts were driven and concrete dams installed in 
various underground tunnels to prevent the flow of the solution into 
active mining areas on the fourth level. It was also necessary to in- 
stall on the third level a concrete ditch with a capacity of 500 gallons 
per minute, together with the necessary pump station and pumps for 
handling the water. 

An underground pumping station delivers the solution to the surface 
through an 8-inch lead-lined pipe. Two centrifugal pumps made of 
Duraloy are used; these have a combined capacity of about 500 gallons 
per minute. 

Water is pumped to the caved area by means of a four-stage cen- 
trifugal pump with a capacity of 340 gallons per minute; another 
pump can increase this supply to 500 gallons per minute when the water 
is available. The water was originally distributed over the caved area 
by a system of pipes equipped with rotating sprinklers, but it was 
found that when the water was allowed to run too long in one place, 
channels were formed, and the water ran through these without dis- 
solving much copper. At present the sprinklers are placed in one sec- 
tion and allowed to remain there until the copper content of the liquid 
drops to 0.4 per cent; then they are moved to another section of the 
caved area. The caved area can be reworked repeatedly because the 
draining period between sprayings permits the old channels to seal 
themselves. The alternate periods of spraying and draining aid in the 
oxidation of the pyrite to ferric sulf ate the principal solvent for the 
copper minerals. 

Leaching has been conducted entirely by fresh water, although it 
may be necessary to employ a leaching agent at some future time. 
Possibly the tailing water from the precipitation plant can be used for 
this purpose, taking advantage of the ferric sulfate formed by the oxi- 



RAY MINES 



337 



dation of ferrous sulfate. This tailing water would require condition- 
ing, however, before it could be used as a leaching agent because when 
ferrous sulfate is oxidized a basic iron precipitate is formed, and this 
would tend to seal up the channels through the broken ore. 

Figure 8 shows the general plan of the precipitating plant which was 
designed after considering the advantages and disadvantages of a 
number of other installations for recovering copper from mine waters. 
The precipitation is carried out in two sections each containing 5 cells 
10 feet 8 inches wide and 40 feet long with an 8-inch dividing wall 




Railroad spur track for detinned iron scrap 



Working runway for crane 
Inlet section 2 Outlet section 2 , Outlet section 1 Inlet section 1 

U .^ u.1 1 c'o" s i_J 




Ml 


V- VU, 2^-s' N. 






Storage area for 
detinned iron 










i-i 


LJE 


XJ 


a-i 




_i 


IJE: 






V s 


*U 
g 


Y 

o' 


I 


T 

Storage area for 
detinned iron 




=5 




r= 


=5 




= 


= 


=1 


= 


J 


= 


=f 


: 


BC 


= 




^ 






Return 
























, 


^ 




^ 


. 


o 

h 


I 




w * crllnc > Oainotlets. $3 


'1 ' " 


-JjL ' " -' - J " " " s 










|De"pth 3 r O' 
li 




4'^10 ( 




.10" 


P m "^ "jh> 


"]" 








CO 






Wash water MU 
return sump jj 


Storage 
cell 


Drying 
cell 


Drying Drying 
cell cell 


Drying 
cell 


Dryi]ng 


Storage 
cell 




Auxiliary drain 


* * r7 r R ^ Depth 3'6" T 10" 





Electric 
tower 



Loading runway for crane 



Railroad spur track for loading precipitates 

(TAomair, Mining and Metallurgy, Vol 19, p 48 f, 1938) 

FIG. 8. General Plan of the Precipitating Plant at Ray Mines. 

down the middle. The cells are constructed of concrete and each cell 
is poured as a unit to avoid any cracks which might result in leaks. 
The cells have sloping bottoms, and at the lowest point is a discharge 
opening closed by a special lead discharge valve. Corrosion of the 
concrete by the copper sulfate is noticeable but not serious. 

The scrap iron used is shredded iron made from detinned cans. This 
is loaded into the tanks by means of a %-yard clamshell bucket op- 
erated by a crane. Solution flows through the cells, and when the 
cement copper is to be removed a cell can be cut out and the copper 



338 HYDROMETALLURGY 

flushed out. These cells contain false bottoms consisting of wooden 
grilles upon which the scrap iron rests; the cement copper is washed 
through the grille and into the drying tank by means of an ordinary fire 
nozzle with 125-pound water pressure. Each flushing yields from 
20,000 to 25,000 pounds of copper; and the operation takes about 45 
minutes and requires 4500 gallons of water for each cell. 

The copper precipitate settles in the first drying cell, and the clear 
water is drained off and returned to the precipitating cells. The 
cement copper is transferred to two other drying cells before it finally 
reaches the storage cell from which it is shipped. The rehandling of 
the precipitate is the major factor in eliminating the moisture content. 

The plant is designed to handle 500 gallons per minute; the plant 
data for the period January 29, 1937, to July 1, 1938, are given in 
Table 5. 

TABLE 5 
LEACHING OPERATIONS AT RAY, JANUARY 20, 1937, TO JULY 1, 1938 

Per cent copper in leach solution 923 
Per cent copper in precipitation plant tailing water 0079 

Indicated recovery, per cent 99 14 

Copper produced, pounds 10,201,364 

Scrap consumed, pounds 11,739,340 

Ratio of scrap consumed to copper produced I 15 

Moisture content of precipitate, per cent 22 78 

Copper content of dry precipitate, per cent 87 266 

Kennecott. 13 The ammonia leaching plant of Kennecott, Alaska, 
was put in operation in 1916 and has operated until recently. This 
will serve as an illustrative example of an ammonia leaching operation. 

The Kennecott ore contained both sulfides (chalcocite with some 
covellite) and copper carbonates (malachite and azurite) in limestone- 
dolomite gangue. The sulfides were removed by gravity concentration 
and flotation and the carbonates by leaching; ammonia leaching was 
used because the acid-soluble gangue precluded the use of an acid 
leach. 

There are two essential differences between a plant such as this for 
copper carbonates and a leaching plant such as the Calumet and Hecla 
for native copper: (1) The necessary C0 2 for the formation of am- 
monium carbonate is supplied by the ore in carbonate leaching, but 
in leaching native copper the C0 2 must be provided from some other 

13 Duggan, E J., Ammonia Leaching at Kennecott: Am. Inst. Min. & Met. Eng. 
Trans., Vol. 106, p. 547, 1933. 



KENNECOTT 339 

source; and (2) the solution must be passed through aerating 
towers when native copper is being leached, to provide the neces- 
sary oxygen, whereas in leaching carbonates this latter step is not 
necessary. 

Leaching Feed. The feed to the leaching plant consisted of con- 
centrator tailings containing about 1 per cent copper, mostly in the 
form of carbonates. Fine sand and slime were treated by flotation, 
and the leaching feed consisted of fairly coarse sands (about 20 mesh). 
This material was dewatered to about 5 per cent moisture in Esperanza 
classifiers and charged into a storage bin; from here it was conveyed to 
the leaching tanks by bucket elevators and horizontal chain-drag con- 
veyors. 

Leaching Tanks. Eight leaching tanks were used, each 30 feet in 
diameter; four of these held a charge 15 feet in depth (460 tons) and 
the other four held a charge 20 feet deep (575 tons). The leaching 
tanks were of all steel construction and vapor tight; a false bottom of 
screen was placed about 4 inches from the bottom of the tank, on 
which was placed a filtering medium of coco matting and 8-ounce duck. 
Mechanical excavators were used to remove the tailing after leaching, 
and when run in reverse these served as distributors. Feed was intro- 
duced through a door in the center of the top of the tank, and leached 
tailing was discharged through doors in the bottom to a system of 
conveyor belts which transported the tailings to the dump. 

As in all sand leaching processes, the leaching operation was a batch 
process as far as the ore was concerned; a tank was filled with ore, 
leached and washed, and then the tailings were excavated and carried 
to the dump. 

Chemistry of the Process. The solvent was ammonium carbonate 
containing a slight excess of ammonia, the leach liquors usually con- 
tained about 20 per cent more C0 2 than NH ;i by weight, which 
means a slight excess of NH :i over that required by the formula 
(NH 4 ) 2 CO ;J . Ammonia was purchased as aqua ammonia contain- 
ing about 30 per cent NII 3 , and the C0 2 came from the dissolution of 
copper carbonates. Tins produced an excess of C0 2 , and part of this 
gas was wasted in the evaporators. 

The ore was leached by percolation in the usual manner, the copper 
carbonates going into solution to produce cupric ammonium car- 
bonates. There was no fouling of the solution or building up of 
deleterious salts. After the enriched copper pmmonia solution was 
withdrawn, the ore was washed first with a weak solution and then with 
live steam. The pregnant solution was boiled with steam to drive off 
the NH 3 and C0 2 and to precipitate copper as black oxide (CuO) . The 



340 



HYDROMETALLURGY 




KENNECOTT 



341 



precipitate was filtered from the waste liquor and shipped; the am- 
monia was condensed and returned to the leaching circuit. 

Leaching Process. Figures 9 and 10 show two flowsheets of the 
leaching plant. Figure 10 shows the sand flow through the plant and 
the relative positions of the various evaporators, condensers, and stor- 
age tanks. The " sand flow " is simple the sand is merely charged 
into a tank, leached and washed, and then removed as tailing. 



o o 

"O 



Storage 




Wash 
Solution 



Condenser OC^'O 

Absorber 







Evaporator Evaporator 


Condenser 0(*"v 





Absorber 


Evaporator Evaporator 


Condenser Qfc-Dj 







Evaporator Evaporator 


Absorber 







O 


Make up 


[\ Rich Sol. 


$ (7} 
\1J 


Stor, fl e ^\ 


NH.Conc. 


Extra 


nd 
Make-up 


Make-up 








(Ditggan, Am Inst Mm and Mel. Eng Trans., Vol 106, p 555, 1933) 

FIG. 10. Sand " Plow " for Ammonia Leaching Plant, Kennecott Copper 

Corporation. 



1. 300-ton storage bin. 

2 16-inch elevator 

3. Drag distributors, one 50 ft long, two 100 ft 

long 

4 Four 575-ton leaching tanks 

5. Four 460-ton leaching tanks 

6. Two 20-inch conveyors. 



7. 20-inch conveyor. 

8 Dragline scraper for tails. 

9 Eight 2-inch centrifugal pumps for solution, 
one 4-inch centrifugal pump for rich solution, 
and one 3-inch centrifugal pump for circulating 
water. 



The flow of solution is shown in Figure 9. The leaching was done in 
two steps. The " first leach " solution was a copper ammonia solu- 
tion from another charge; this was a " still " leach, and the solution, 
which was not circulated, yielded the " rich ; or pregnant solution 
containing the maximum amount of dissolved copper ; this solution then 
went to precipitation. After the first leach, the ore was leached with 



342 



HYDROMETALLURGY 



strong ammonia to extract the rest of the copper; liquor from this leach 
then served for the first leach on the next tank of fresh ore. The 
second leach was a " circulation leach " in which the liquid was 
pumped to the top of the tank and allowed to flow through the ore 
column by downward percolation. After the second leach, the tailing 
was washed with water and steam and then discharged. 

TABLE 6 
TYPICAL OPERATING CYCLE FOR ONE CHARGE, KENNECOTT, ALASKA 



Hours 


Operations 


Amount of 
Solution 
(tons) 


NH 3 

(per 
cent) 


Cu 
(per 
cent) 


C0 2 

(per 
cent) 


0- 12 


Charge 575 tons ore @ . 94 per cent 












carbonate Cu, 5 per cent HoO 










12- 24 


Draining 










24- 32 


Pump to bottom first leach solution 


192 


10 50 


6 30 


12.40 


30- 31 


Siphon from top to waste 


10 


trace 


trace 




31- 32 


Siphon from top to wash storage 


6 


2.00 


1.40 




32- 44 


Still leach 










44- 46 


Pump from bottom to rich solution 












storage 


59 


10 32 


7 85 


12.20 


44- 46 


Pump to top concentrated ammonia 


23 


26 00 




22.00 


44- 46 


Pump to top make-up 


36 


4.00 


2.50 


4.30 


46- 94 


Circulation leach 










94-102 


Pump from bottom to next charge 


192 


10 50 


6.30 


12.40 


94- 96 


Pump to top wash 


30 


1.50 


1.00 




98-108 


Steam wash, 28 tons steam to top 










102-104 


Pump from bottom to make-up 












storage 


36 


4 00 


2.50 


4.30 


104-108 


Pump from bottom to wash storage 


24 


1 40 


65 




108-120 


Discharge tailing, . 14 per cent car- 












bonate Cu 











Table 6 shows a typical operating cycle for one charge of ore. The 
total cycle took about 5 days (120 hours). The first 24 hours was 
consumed in charging and draining the ore. The first leach solution 
was then pumped in from the bottom, and the residual moisture was 
siphoned off at the top as it was displaced by the rising column of 
solution. This displaced liquid amounted to about 16 tons; the first 
portion (10 tons) contained only a trace of copper and ammonia and 
was discarded; the last 6 tons contained some copper and ammonia, 
and this was sent to wash-solution storage. The first leach stood on 
the ore for about 12 hours without circulation, and at the end of this 
time the solution would be about three-fourths saturated with copper 



KENNECOTT 343 

: Cu =- 1.3 : 1), which is about the practical limit for good ex- 
traction. A part of this liquid (59 tons in this example) was then 
pumped from the bottom of the tank to the rich solution storage. 
Sufficient strong ammonia (from the condensers plus new ammonia) 
and make-up solution to extract the rest of the copper was then 
pumped in on top, and this " second leach " solution was then circu- 
lated through the column for 36 to 48 hours. At the end of this time 
the liquid was pumped from the bottom of the tank to another tank 
of fresh ore, where it became the " first leach." 

As the second leach was pumped off it was immediately followed by 
20 to 30 tons of wash solution added at the top, and when the level of 
this wash solution was within a few feet of the bottom of the tank, live 
steam was admitted at the top and steaming was continued until the 
washing was completed. The first wash solutions from the tank con- 
tained the most copper and ammonia and went to make-up storage; 
as the steaming continued, the effluent liquid became leaner in copper 
and amtnonia and was sent to wash solution storage. Steaming was 
continued until the NH 3 content of the effluent had been lowered to 
the proper amount, and the tailing was then excavated and sent to the 
dump. 

Originally the steam was forced through the bed at 4 to 5 pounds 
pressure, but later a vacuum was used according to the practice at the 
Calumet and Hecla plant; this innovation shortened the time for the 
steam wash and reduced the steam consumption about 10 per cent. 
Steam washing is essential in the treatment of low-grade material in 
order to secure economical recovery of the ammonia; ammonia ab- 
sorbs strongly on the surface of cold ore and is difficult to wash out 
with water. One ton of steam after the weak solution wash washes 20 
tons of ore without the use of any other fresh water. 

Recovery of the Solvent. The rich solution from storage went to 
the evaporators, where it was boiled with steam until the NH 3 and 
C0 2 were removed and the copper precipitated as CuO (Fig. 9). This 
was essentially a batch process also, as a given amount of rich solution 
placed in one of the evaporators remained there until it was completely 
" boiled out." The evaporators were arranged in a number of (one 
to four) " effects." (See Fig. 9 for a " three-effect " arrangement.) 

This means that the rich solution from storage was first heated by 
vapor which had passed through two other evaporators; after a while 
the live steam was turned on the second tank, pnd the tank which had 
been on " third effect " became " second effect." After this the live 
steam was turned directly into the tank, giving the " first-effect " 
treatment. 



344 



HYDROMETALLURGY 



Vapor from the third-effect tanks passed through a preheater where 
some of its heat was used to preheat the rich solution going to the evap- 
orators, and then it passed to the condenser where it condensed as 
strong ammonia and passed back into the leaching circuit. The first 
gases to come off the third-effect tanks were largely C0 2 , and these 
passed through the condenser to a scrubber which removed any NH 3 . 
The C0 2 gas was then discharged into the atmosphere. The amount of 
C0 2 thus wasted corresponded to the amount of C0 2 picked up by the 
leaching of the copper carbonates. 

Steaming was continued until the liquid on the first-effect tank 
contained only 0.01 per cent NH S ; by this time the copper had all 
been converted to CuO. The liquid was then forced out by steam 
pressure through a filter located below the evaporator; the CuO pre- 
cipitate was dried by vacuum and shipped, and the barren solution 
passed through a sump to collect any suspended precipitate, and on 
to the discard. 

The evaporators were 16-foot steel cylinders 10 feet in diameter 
with dome-shaped tops and 6-foot cones on the bottom. Below each 
evaporator was a filter which served to catch the precipitated CuO. 

Extraction, Etc. The strength of the leach solutions was from 6 
to 11 per cent NH 3 with corresponding copper percentages of 4.5 to 8.0. 
The Cu, NH 3 , and C0 2 content of typical solutions is shown in Table 6, 
together with tonnages of these solutions. The plant had a capacity of 
800 tons of concentrator tailing per day. 

The precipitated black copper oxide contained about 75 per cent 
copper. Little ammonia was lost by volatilization to the atmosphere, 
and the total ammonia consumption amounted to 0.45 to 0.60 pound 
of NH 3 per ton of ore leached. Steam consumption was 210 to 230 
pounds per ton of ore leached; 55 per cent of the steam was used for 
evaporation and 45 per cent for the steam wash. Recovery of copper 
ranged from 88 per cent to 76 per cent, the recovery being greatest 
in the finer sizes (Table 7). 

TABLE 7 
AVERAGE EXTRACTION OF COPPER, KENNECOTT LE ACHING PLANT 



Size of 
Ore Particles 
(mm) 


Carbonate Cop- 
per in Heads 
(per cent) 


Carbonate Cop- 
per in Tails 
(per cent) 


Extraction 
(per cent) 


13 
11 
9 

7 


0.90 
0.90 
90 
0.90 


22 
0.18 
0.14 
0.11 


75.9 
80 4 
84.8 
88.1 



CHUQUICAMATA 



345 



Chuquicamata. 14 The leaching plant at Chuquicamata, Chile, is 
one of the largest plants in the world, having a capacity of 1,400,000 




(Campbell, Am Inst Mm. and Met Eng Trans , Vol 106, p 607, 1933) 

Fia. 11 General Plan of Plant, Chuquicamata. 

short tons of ore per month, which means a production of 20,000 to 
28,000 tons of copper per month The following quotation is taken 
directly from Campbell's article. 

Copper is extracted from the Chuquicamata oxide ore by a hydrometal- 
lurgical process. The ore is crushed to %-inch sizing, and leached with a 
sulfunc acid electrolyte. Chlorine is precipitated and the ferric iron 
reduced in the enriched electrohte, after \\lnch the copper is recovered by 
electrolysis with insoluble Chilex and lead-antimony anodes, the spent elec- 
trolyte being returned to leaching. Cathodes are melted and refined in 
market furnaces and cast into commercial wire-bar and cake shapes. Sul- 
func acid for the process is supplied by the brochantite in the ore. Water 
for washing the ore is advanced through the solution system and finally 
after cutting by electrohsis to from 6 to 16 grams per liter is completely 
stripped by the cuprous chloride method and run to waste. The cuprous 
chloride so obtained plus that resulting from precipitation of chlorine from 
strong solution is dissolved in ferrous chloride brine and the copper cemented 
on scrap iron. The cement so obtained is in part used to reduce ferric 
iron in electrolyte and as it is of exceptional purity is also furnace refined 
and cast into an exceptional quality of fire-refined copper. * * * 

The reduction plant is composed of seven divisions as follows: (1) crush- 
ing, (2) leaching, (3) tailings disposal, (4) dcchlondizing, (5) sulfur dioxide 
treatment, (6) electrolytic tank house, (7) smelting and melting. * * * 

Ore. Table 8 shows the average analysis of the ore and the amount 
of the various constituents extracted by the teaching process. The 

14 Campbell, T. C., A Brief Description of the Reduction Plant of the Chile 
Exploration Company at Chuquicamata, Chile, S. A : Am. Inst. Min. & Met. Eng. 
Trans., Vol. 106, p. 559, 1933. 



346 



HYDROMETALLURGY 



present ore (1933) contains 90 per cent oxide copper and 10 per cent 
sulfide copper; 98 per cent of the oxide copper and 40 per cent of the 
sulfide copper are extracted by the normal leach. 

A moderate tonnage of " border-line mixed ore " is leached with 
large volumes of solutions containing ferric iron; these ores contain 
60 per cent oxide and 40 per cent sulfide copper. As high as 70 per 
cent of the sulfide copper in this mixed ore can be extracted. When 
ferric sulfate is reduced by sulfides in the ore, an equivalent tonnage 
of scrap iron is saved in the dechloridizing operation. 

TABLE 8 
ORE ANALYSIS AND EXTRACTION AT CHUQUICAMATA 





Analysis 


Extraction 


Per Cent 


Kilos per 
Metric Ton 
of Ore 


Kilos Ex- 
tracted per 
Metric Ton of 
Ore 


Extraction, 
Per Cent 


Copper 
Iron 


2.100 
0.95 


21.0 
9 5 


19 5 
0.15 


23.0 
15 8 


Chlorine 


012 


0.12 


0.04 


33.3 


Arsenic 


0.002 


0.02 


01 


50 


Antimony 
Molybdenum 
Nitric acid 


0.001 
0.010 
0.015 


0.01 
0.10 
0.15 


0.004 
0.01 
0.08 


40 
10 
53.3 


Sodium 


1.10 


.... 





.... 


Potassium 


5.20 








Silica 


68.0 








A1 2 8 


16 









The oxide ore is a highly altered granite containing veinlets of 
brochantite with minor amounts of other oxide minerals; the sulfide 
minerals are chalcocite and pyrite, and very little of the pyrite is 
attacked by the leach solutions. 

The ore is mined by open-cut methods and when delivered to the 
crushing plant contains pieces as large as 5 feet in diameter. This is 
crushed down to a 0.371 -inch sizing by four-stage crushing. Two 
60-inch Superior McCully gyratory breakers reduce the ore to 9-inch 
diameter. In the second stage the 9-inch material is reduced to 3-inch 
sizing in seven No. 10 McCully gyratories and one No. 7 cone crusher. 
The third and fourth stages of crushing are accomplished by fifty 
48-inch disk crushers; 14 of these take the ore from 3-inch diameter 
down to 1-inch, and the remaining 36 disk crushers take the 1-inch 
material and turn out the final product of -0.371 inch. The crushed 



CHUQUICAMATA 347 

ore is taken by conveyor belts to the loading bridges over the leaching 
tanks. 

The brochantite supplies about 0.40 kilo of sulfuric acid per kilo 
of copper leached from the ore; this is sufficient acid to operate the 
plant, and no other acid source is required. 

Leaching Equipment. The ore is leached in 13 leaching vats, each 
150 feet long by 110 feet wide and 16% to 18% feet deep to the top 
of the filter bottom, with a net capacity of 11,500 tons of ore in each 
tank. The tanks, constructed of reinforced concrete, are arranged in 
blocks of three or four and are set on piers to facilitate inspection of 
their bottoms. The tanks have a 4-inch lining of a mastic sand 
mixture. 

On this mastic floor a false bottom is laid composed of 6- by 6-inch 
pine 6 feet long laid end to end 18 inches apart. Across these are 
laid 2- by 6-inch pine planks spaced % inch apart. Eight spaces 10 
feet square are left in this false bottom for draining the tank; into 
these are inserted prebuilt filter units composed of 4- by 6-inch timbers 
laid 12 inches apart, across which are laid 2- by 6-inch planks set 3 
inches apart. Over this is laid coco matting, and on the matting 
another layer of 2- by 6-inch planks % inch apart is laid with the 
planks parallel to the closing movement of the excavating buckets. 
In excavating the tailings a small tonnage (800 tons) is left on the 
filter bottom to prevent injury to the filter. This filter bottom will 
drain 1500 cubic meters (400,000 gallons) of solution an hour through 
average ore and with occasional minor repairs will last about 8 years. 

Leaching vats are arranged in two units, one of seven vats, and the 
other of six (Fig. 11). Each unit is served by one loading and two 
excavating bridges of the gantry type; the excavating bridges span 
the loading bridge. Two main loading conveyors run the length of 
the north side of the two vat units, and four tailings trains operate 
on the south side taking the discharge from the four excavating bridges. 

The ore is loaded either into solution or a dry vat, depending upon 
the production cycle. Ore is not bedded but is loaded from bottom 
to top before the bridge is advanced. The loading bridge is advanced 
into the prevailing wind to prevent a layer of wind-blown fines from 
covering the exposed filter bottom, and curtains are suspended from 
the bridge to shield the ore from the wind as much as possible. After 
loading, the ore is leveled by hand and fines on the surface are 
turned in. 

Tailings are excavated by means of clamshell buckets operated from 
the excavating bridges; these buckets have capacities of 8 to 12 tons. 
Two bridges operating together on a single vat can excavate 11,000 



348 HYDROMETALLURGY 

tons of tailings in 7 to 7 l / 2 hours. Tailings discharge into hoppers 
from which they are hauled in 12-cubic-yard dump cars in trains of 
26 to the tailing dump. 

There are 25 solution sumps for " active " solutions and four auxiliary 
storage (" passive storage ") sumps. These have a combined capacity 
of 75,000 cubic meters (20,000,000 gallons). The sumps are placed 
on higher ground than the leaching vats so that solution can be 
run into the vats by gravity. The sumps are used for transient 
storage of treatment advance and wash solutions and as buffer tanks 
for spent and strong electrolyte. Construction of the sumps is similar 
to that of the leaching vats but is somewhat lighter. 

The leaching and electrolytic plants contain 68,865 feet of pipe, 
with 885 valves and 2090 elbows or tees, for transmitting solutions. 
Pipes are of 8-, 12-, 15-, and 24-inch diameter, depending upon the 
service, and all pipe is wood-stave pipe made of Oregon pine and 
Douglas fir. Pipe is made in 17-foot lengths, and sections are 
coupled with wood-stave couplings. Pipe fittings are of cast iron 
with mastic or hard lead linings. The average life of wood-stave pipe 
is 5% years. 

Nineteen 15-inch and four 9-inch vertical centrifugal pumps are 
employed to circulate the solutions. These have runners, casings, 
and boots all covered with hard lead, and the pump intakes are con- 
nected to mastic-lined sumps equipped with cast-lead screens. As 
high as 250,000 cubic meters (66,000,000 gallons) can be pumped daily 
against a 70-foot equivalent head. 

Leaching Method. The method used for leaching is the batch perco- 
lation system in distinction to the countercurrent percolation system 
used elsewhere. By this treatment each tank is leached more or less 
independently of the others; the countercurrent system gives better 
recovery, but the batch system is more flexible. 

Approximately 3500 cubic meters (925,000 gallons) of solution is 
required to submerge 10,000 metric tons of average ore; of this 750 
cubic meters is absorbed and 2750 cubic meters is drainable. 

The flow of the various solutions through the leaching and electro- 
lytic cycle is shown in Figure 12. The flow is complex, and the 
details of leaching will vary from time to time, depending upon the 
grade of the ore and the production rate desired. In general, the 
purpose is to leach fresh ore with a partly spent solution to produce a 
strong (high copper) solution, which goes to the electrolytic plant. The 
partly leached ore is then treated with a strong acid solution to com- 
plete the removal of the copper; this solution when withdrawn then 
goes on another tank of fresh ore, where it picks up enough copper to 



CHUQUICAMATA 349 

go out as strong solution. A certain amount of leach solution is 
discarded. 

After a tank is charged and leveled the remainder of the treatment 
solution is run on through the bottom until the ore is submerged. 
The ore is then allowed to soak from 8 to 24 hours. Then the pro- 
duction of " first strong solution " is started from the bottom of the 
tank with a solution called " first advance " going on top at the 
same time; first strong solution is produced to a " cut-off " limit (i.e., 
solution is withdrawn until the copper content reaches a certain 
minimum value). The volume of first strong solution ranges from 
3000 to 6000 cubic meters, depending upon the grade of the ore. 
Following the production of first strong solution, a second soaking 
period ensues of 24 to 72 hours, after which the production of " second 
strong solution " starts, with " second advance " solution going on top. 
Second strong solution is also produced to a cut-off limit, and the first 
strong solution from one tank is blended with the second strong solution 
from another tank to make the solution which goes to the electrolytic 
tank house. 

After these two steps there are three more soaking periods and six 
washes, as follows: 

At the end of the third soaking period, which is of variable duration, 
" first advance " solution is produced and " volume advance " solution 
and " spent electrolyte " go on top. 

At the end of the fourth soaking period, which is also of variable 
duration, " second advance " solution is withdrawn, with spent electro- 
lyte going on top. 

At the end of the fifth and final soaking period the washing process 
starts. It continues until the ore is completely washed. Then the 
ore is drained for at least 4 hours and excavated. 

The solutions going on top of the tanks after the fifth soaking are 
six washes of decreasing copper grade ranging from 9 grams to 1 gram 
per liter, followed by a water wash. The solutions drawn from the 
bottom of the tank during this period are determined by cut-off limits, 
and in order of their production are: " treatment " or " covering solu- 
tion," " volume advance/' the six wash solutions, and " volume discard." 
The " volume advance " solution is made in quantity sufficient to re- 
place the volume loss of the leaching solution and is equal to the 
volume loss of primary spent electrolyte stripped and wasted plus the 
evaporation loss in the primary leaching system. The "volume dis- 
card " solution is the same as the " volume advance " solution. It is 
sent through the dechloridizing plant and to the discard. 

The distinction between " volume advance " and " volume discard " 



350 HYDROMETALLURGY 

depends upon what use is made of the solution. No new water enters 
the leaching system, but some is lost by evaporation and discarding 
of electrolyte this is made up by " volume advance " solution. In 
the washing system, however, new water enters the system in excess 
of the amount lost by evaporation and entrained in the tailings. This 
excess is the amount of " volume discard." 

Table 9 gives the average composition and tonnages of the various 
solutions involved in leaching a vat of ore. 

Ninety-six hours is the minimum cycle for good leaching. Leaching 
operations are scheduled on a time chart days in advance to prevent 
conflicts and to give a close control of operations. 

Dechloridizing Plant. The dechloridizing plant is an essential part 
of the plant and it performs other functions besides the removal of 
chlorine from strong solution. The five important functions of the 
dechloridizing plant are: 

1. Precipitation of chlorine from strong solution as cuprous chloride. 

2. Reduction of ferric iron in strong solution. 

3. Stripping of copper from electrolyte to be wasted. 

4. Recovery of cement copper from cuprous chloride by redissolving 
in hot ferrous chloride brine and cementing on scrap iron. 

5. Preparation of high-grade cement copper from which a high- 
grade fire-refined copper is made. 

The precipitation of chlorine and reduction of ferric iron by cement 
copper is carried out by agitation of the strong solution with cement 
copper, and the reactions are the same as those given on page 328. 
Cuprous chloride precipitate is dissolved in FeCl 2 brine and the 
copper cemented out on scrap iron. Part of the precipitated copper 
is then returned for treatment of more strong solution, and the re- 
mainder goes to the melting furnaces. 

Waste electrolyte can be stripped of its copper content quickly and 
efficiently by the action of cement copper and ferrous chloride. A small 
amount of FeCl 2 solution is added to the electrolyte and is agitated 
with cement copper; the reaction involved is 

Cu + CuSO 4 + FeCl 2 -> 2CuCl + FeSO 4 

The precipitate is settled out for treatment to recover the copper, and 
the liquid is then discarded. A slight excess of FeCl 2 and a large excess 
of cement copper is required to carry this reaction to completion in a 
short time. 

Some chlorine is lost in discarded electrolyte and in washing opera- 
tions, and as the chlorine content of the ore has diminished with the 
depletion of surface ores, the chlorine dissolved from the ore is no 



CHUQUICAMATA 



351 



I 

O 



-I 



1, 

OJ I 



3 .s 



1, 



I 1 

E i 



$ <u *? 

a 






> H c ' 

aoJ 3 
s s ~ 



'S , 8 

O T3 fl 



'O bfi , 

a s i 
8 8-3.2 



(N 



o >o o 

8 8 g - 



b i 

g i c 
P "3 P 



o o 



i> oo o 



"?3 fl 



nT cf 

g a >H 2 ^ 32 

il'lltllli 

&H ^ 



2 



352 HYDROMETALLURGY 

longer sufficient to satisfy the needs of the dechloridizing plant. Ac- 
cordingly, sodium chloride crystallized from nearby desert springs is 
added to make up the required chlorine. At present (1933) about 45 
per cent of the required chlorine comes from the ore, 25 per cent from 
the wash water, and 30 per cent from NaCl. It will undoubtedly prove 
economical and desirable in the future to add NaCl, even though 
the chlorine content of the ore were to drop so low that dechloridizing 
of the strong solution would no longer be necessary. 

Two principal advantages of the chlorine method of stripping waste 
solutions are (1) it is possible to produce a higher grade of cement 
copper than by direct precipitation on metallic iron, and (2) the con- 
sumption of scrap iron is only about half as much as would be required 
for direct precipitation. The CuCl is redissolved in ferrous chloride 
brine and precipitated on scrap iron, so the stripping of the solution is 
actually performed by scrap iron, although indirectly. In the electro- 
lyte, however, the copper is in the cupric form, and twice as much iron 
is required per pound of copper precipitated as is required in precipi- 
tating the cuprous copper in ferrous chloride brine; thus: 

CuSO 4 + Fe -> Cu + FeS0 4 
2CuCl (in FeCl 2 sol.) + Fe -> 2Cu + FeCl 2 

TABLE 10 
DECHLORIDIZING PLANT DATA, CHUQUICAMATA 

Dechloridizing capacity, cubic meters of strong solution per 24 hours 25,000 

Tonnage capacity to cementation, short tons per 24 hours 140 

Normal chlorine in entering strong solution, gram per liter 0.50 

Normal chlorine in leaving strong solution, gram per liter 0.05 

Scrap iron consumption per unit of cuprous copper 50 

Grade of solution to stripping, grams per liter of copper 8 to 40 

Grade of solution from stripping, gram per liter of copper . 4 

Sulfur Dioxide Plant. In 1930 a sulfur-burning plant with S0 2 
absorption tower was installed. The plant burns a local volcanic 
sulfur ore to produce S0 2 , and the S0 2 gas is absorbed by the electro- 
lyte. 

The purpose of this treatment is not to reduce ferric iron, for which 
purpose the process is employed elsewhere, but simply to dissolve S0 2 
in the electrolyte to stabilize it. The contact of solution with S0 2 gas 
is just long enough to permit dissolution of some of the gas and is not 
prolonged enough to allow much reduction of ferric iron. 

Previous to this installation operating difficulties were encountered 
in the tank house by the decomposition of nitric acid ; this was catalyzed 



CIIUQUICAMATA 353 

by molybdenum in the solution when ore containing over 0.008 per 
cent molybdenum was treated. Once started, the reaction was self- 
catalyzed by nitrogen oxides and spread through the entire solution 
system of 100,000 cubic meters. Dechloridizing and stripping opera- 
tions were hampered, and the chlorine content of the electrolyte and 
cuprous chloride in the cathodes increased greatly; current efficiency 
dropped from 90 to 60 per cent; and the anodes and pipe fittings were 
severely corroded. 

Experimental work indicated that a small amount of sulfurous acid 
(S0 2 + H 2 O > H 2 S0 3 ) m solution would inhibit the decomposition of 
nitric acid and would gradually dissipate the nitnc acid without oxida- 
tion of anodes and fittings. Accordingly the SO 2 plant was constructed, 
and the flowsheet was arranged to keep S0 2 dissolved in the electro- 
lyte by continually circulating a portion of the electrolyte through the 
SO 2 absorption plant. 

The plant burns 25 tons of fine sulfur per day producing about 
50 tons of S0 2 in a gas of 14 per cent grade. The gassed electrolyte 
carries 0.7 gram to 2 grams of S0 2 per liter, and with electrolyte 
containing 0.2 to 5 gram of S0 2 per liter there is practically no odor 
of S0 2 in the tank house; escape of S0 2 is largely prevented by the oil 
blanket which covers the electrolyte in the cells. About 90 per cent 
of the absorbed SO 2 is converted to sulfuric acid by anodic oxidation 
in the cells. 

Electrolysis. There are 1098 electrolytic tanks each measuring 19 
feet 2 inches long, 3 feet 11 inches wide, and 4 feet 10 inches deep, in- 
side dimensions. The tanks are of reinforced concrete and are lined 
with mastic. The deposition tanks rest on piers with ground footings 
constructed independently of the building proper, and the tank tops 
are 2 to 3 feet above floor level. 

A section comprises sixteen or seventeen tanks arranged in a cascade, 
and four to eight sections in series make up an electrical circuit. Each 
circuit is powered by one or more rotary converters or motor-generator 
sets. One circuit and a variable number of tanks in another circuit are 
used for the deposition of starting sheets, and one or more circuits 
are always in use plating down solution which is to be discarded. The 
remainder of the tanks are used for commercial cathode deposition. 

The anode plant adjoins the tank house, and here the copper 
silicide (Chilex) anodes and lead-antimony grid anodes are cast. 
Chilex anodes are 2 feet 9 inches by 5 feet 11 irches, are 1 inch thick, 
and are spaced on 3%- or 4-inch centers. Lead-antimony anodes are 
of the same width and length but are 0.5 to 0.6 inch in thickness and are 
spaced on 3- to 3% -inch centers. Each tank, therefore, carries 56 



364 



HYDROMETALLURGY 



to 74 anodes and 55 to 73 cathodes, depending on the spacing. Chilex 
anodes are brittle, and the lead-antimony anode produces about 12 



SPENT ELECTROLYTI 

OR OTHER DISCARD 

SOLUTION 




REDUCTION OF 
FERRIC (ROM IN 
STRONG SOLUTION 




OECHLORIDIZINQ 
PLANT 


HIGH GRADE 
COARSE CEMENT 



MARKET 
rURNACE 



COMMERCIAL SHAPES 




(Campbell, Am. In*. Min. and Met. Eng. Trans., Vol. 106, p. 694, 1983) 
FIG. 13. Flowsheet of Electrolytic Tank House, Chuquicamata, 1933. 

per cent more copper per kilowatt day than the Chilex anode. The 
Chilex anode resists corrosion better than the lead-antimony anode, but 
with the low iron and nitric acid content of present-day electrolyte, 



CHUQUICAMATA 355 

this is not such an important factor as it used to be, and lead-antimony 
anodes are gradually replacing the Chilex anodes. The tank house 
requires from 60,000 to 62,000 anodes. 

Starting sheets are made from dechloridized strong solution using 
lead-antimony anodes. Equipment is also provided for making starting 
sheets from soluble anodes (made by fire refining cement copper) 
with an electrolyte made by dissolving bluestone. The method used 
for making starting sheets depends on the condition of the strong 
solution. If more than 0.2 gram per liter of chlorine is present the 
starting sheets are too brittle for use. The sheets are deposited to 
12-pound weight and the submerged section measures 3 by 4 feet. 
Two loops cut from the starting sheets attach them to the cathode bar. 

The outflowing solution from the stripper tanks is pumped to the 
S0 2 plant and then enters the commercial cells. The flow of the 
electrolyte is very fast in order to insure formation of hard cathodes 
and to decrease polarization. Each section takes a flow of 750 liters 
(198 gallons) per minute, and at times the flow may be as high as 
1200 liters (317 gallons) per minute. The electrolyte is not heated 
except by the cell resistance, which causes a rise of about 10 C as the 
electrolyte passes through the cells. The temperature of the spent 
electrolyte will be 30 to 45 C, depending on atmospheric conditions. 
The average weight of a finished cathode will be about 150 pounds, 
production of which has required 5 to 15 days, depending on the current 
density. 

TABLE 11 
ELECTROLYTIC TANK HOUSE DATA, CHUQUICAMATA 

Number of electrolytic tanks 1,098 

Number of insoluble grid anodes to equip tanks 62,000 

Current density, amperes per square foot of cathode area 7 to 18 
Entering solution (Temp. 26-34 C), grams per liter: 

Copper 21 to 26 

Ferrous iron 1 . 6 to 2 . 1 

Total iron 2.5 

Leaving solution (Temp. 31-43 C), grams per liter: 

Copper 14 to 16 

Ferrous iron . 5 

Total iron 2.5 

Current efficiency 85 to 92% 

Pounds of copper per kilowatt-day 24 to 28 

Capacity, kilowatt load 55,000 

Current density ranges from 7 to 15 amperes per square foot of 
cathode in the stripper cells and from 7 to 18 amperes in the com- 



356 HYDROMETALLURGY 

mercial cells. Voltage drop per tank is about 1.9 to 2.0 volts for 
lead-antimony anodes and 2.1 to 2 3 volts for Chilex anodes. 

Smelter. The smelter has three market furnaces (cathode furnaces) 
of 400 tons daily capacity each for melting cathodes and casting wire- 
bars, cakes, and other shapes. Each furnace is equipped with a 
40-foot Clark casting wheel. 

There is a market furnace (reverberatory furnace) of 150 tons 
capacity for the fire refining of cement copper; the technique employed 
is similar to the process used for smelting native copper. There is 
also a small blast furnace of 50 tons daily capacity for smelting re- 
finery slags and miscellaneous secondaries. When soluble anodes 
are made, the black copper from the blast furnace is refined in the 
150-ton reverberatory and cast into anodes. 

Inspiration. 15 Practice at the Inspiration plant of the Inspiration 
Consolidated Copper Company at Inspiration, Arizona, differs from 
that at Chuquicamata in several important respects. 

1. The ore is not suited for all-sand leaching, and part of it is 
treated by percolation and part by agitation. 

2. The amount of sulfide copper is greater, and more of the sulfide 
copper is dissolved by ferric sulfate than at Chuquicamata. 

3. It is not necessary to dechloridize the solution; and* no effort is 
made to reduce ferric iron, as this is an effective leaching agent. 

4. The ore does not supply sulfuric acid, so new acid must be con- 
tinually added to the system. 

The Ore. ^hc copper content of the ore is about 1.3 per cent, of 
which about 0.7 per cent is sulfide copper (1931). The fluctuation of 
ore and tailing assays in the period 1927 to 1931 is shown in Figure 14. 

Ore is crushed in gyratory crushers and Symons disk crushers which 
reduce mine-run ore to about 1% inches. This is further crushed by 
coarse and fine rolls in closed circuit with a screen with %-inch open- 
ings. Part of the undersize from the crushing plant (Fig. 15) goes to 
the washing plant, and the remainder goes directly to the sand leach. 

The important oxidized copper minerals in the ore are chrysocolla, 
malachite, and azurite. Chalcocite is the principal sulfide mineral, and 
the ore contains very little soluble iron. 

The plant has a yearly capacity of some 3,000,000 tons of ore. 

Washing Plant. The washing plant takes part of the ore from 
the crushing division (the undersize containing the " natural slimes " 
in the ore) and washes it in two 25-foot Dorr bowl classifiers. The 
sands from these classifiers join the rest of the crushed ore and go to 

15 Aldrich, H. W , and Scott, W. G , The Inspiration Leaching Plant: Am. Inst. 
Min. & Met. Eng. Trans , Vol. 106, p. 650, 1933. 



INSPIRATION 



357 



the leaching tanks. The classifier slimes go to the flotation and slime- 
leaching divisions. 

Taking 1931 as an average year, 5.75 per cent of the total ore mined 
was removed and sent to flotation and slime leaching; this material 
assayed 1.545 per cent total 

1.3 r- 



1.2 



1.1 



1.0 



: 0.7 



- 6 



0.5 



0.4 



0.3 



x e Copp, 



.!_, 



copper and 0.389 per cent sul- 
fide copper. The classifier 
sands contained 3 3 per cent 
200-mesh material, and the 
classifier overflow was 83 4 per 
cent 200 mesh. Water used 
per ton of slime amounted to 
1045 gallons. 

The wet classifier sands are 
mixed with the dry crushed ore 
and are sent to the leaching 
vats. 

Sand Leaching Equipment. 
There are 13 concrete, lead- 
lined leaching tanks each 175 
feet long, 67.5 feet wide, and 
18 feet deep, with a capacity of 
9000 tons of dry ore per tank. 
The lead on the side walls is 
protected by a covering of 
2-inch planks held in place by 
vertical posts. The filter bot- 
tom is made of 2-inch boards 
having fifteen %-inch holes per 
square foot of surface, each 
countersunk with a %-inch hole 
from beneath. The tank bottom slopes slightly toward the center 
as well as toward the end where the drain pipe is located. There is 
only one opening in the bottom of each tank a 14-inch lead pipe 
burned to the lead lining, entering at the end opposite the overflow 
just above the tank bottom and below the filter bottom. All solution 
enters through this pipe and all drainage is taken out through it. 
Timbers 4 by 6 inches and laid on the 4-inch side are placed over the 
filter bottom to protect it, and a 6-inch layer of tailing is left when 
excavating. 

The thirteen tanks are arranged in a single row and are served by 
one excavating bridge and a loading bridge. The loading bridge 



0.20 g 
I 
c 

A 0-15 I 



1927 



1928 



1929 



1930 



0.10 



1931 



(Aldrich and Scott, Am. Inst Mm. and Met. Eng. 
Trans , Vol 106, p. 650, 1933) 

FIG. 14 Copper in Feed and Tailing, 
Inspiration Leaching Plant. 



358 



HYDROMETALLURGY 



contains a conveyor belt with an automatically reversing tripper which 
empties the load into the tank. At each reversal of the tripper the 
loading bridge moves forward 2% feet until it reaches the end of 
the tank, when its motion is reversed. This lays the charge in the tank 
in a series of beds each about 3 feet thick. The excavating bridge uses 
an unloading bucket capable of lifting approximately 17 tons of wet 
tailing; two 8-hour shifts are required to unload a tank. 

Leaching Method. The leaching method is simple, being a straight 
countercurrent system using upward percolation, followed by ten 



Ore from 
rse-crushing Plant 



10,000-ton 

Storage 

Bins 




I 

rS Pocket 
U Bin. 

CTO ,, oio y 



/ * * I l.sfn.S. 6-W 5-WI4.W 3.w|2.w]l.W C.S. C.S. 

Iron Launders '..'.. '.. l .'' '-.'.'' . ' 



._. 
4.* Jt L - U- JtA. -* t. _<U_ i_ JL* 

Jj I 1st Wash WateT e 

&1 



'to Dump 



Ula.Jw.'.hX i 1 
ii II 1 1 i 1. i 

Mffir 



To Iron Launders 




13 


12 


11 


10 


9 


B 


7 


6 


5 1 4 


3 


2 


1 




Add 


Treat 


nent 




V 


ashln 
W 


I 
W 


9000 


S 






1-4 


-U4 


4-4 


4-4 


4-4 








It 


-4 


i-t 


4-4 



A =Acid Treatment 
W=Wash!ng 
Ex Excavate 
L = Loading 
I.S. Iron Solution 
C.S.=Copper Solution 

Ore 

, .Solution 



Wash Water - 



(Aldrich and Scott, Am. Inst. Min and Met. Eng. Trans , Vol. 106, p. 668, 19S8) 

FIG. 15. Flowsheet of the Inspiration Leaching Plant. 

washes. The complete cycle is 13 days, and at any one time there 
will be eight tanks on acid leach, three on washing, one being loaded, 
and one being excavated. 

From 175,000 to 200,000 gallons of solution are required to cover a 
tank of ore. Each tank is provided with a vertical screw-type lead 
pump which takes the solution overflowing from the preceding tank 
and pumps it underneath the filter bottom and up through the ore. The 
spent electrolyte from the tank house plus new acid is added to the 
oldest ore on acid leach. The solution travels from tank to tank, being 
constantly reduced in solvent strength and increased in copper content, 
until it emerges from the vat containing the newest ore. From there 
it flows to the tank house. 

Washing requires 3 days, and ten washes are used. The first five 



INSPIRATION 



359 



are called regular washes and are systematically advanced, the first 
wash going to the main solution system, the second becoming the 
first wash on the next tank, and so on. 

The solution for the fifth wash comes from the cementation solution 
stock, and aftei passing through the ore becomes the fourth wash on 
the next tank, and eventually works its way into the main solution 
circuit. The next four washes come from the cementation stock 
solution, and the tenth wash employs fresh water. The washings from 
the last four washes go to cementation. This liquid is passed over 
scrap iron, which precipitates copper, reduces ferric sulfate, neutralizes 
the acid present, and adds ferrous iron to the solution. This latter is 
the principal reason for the cementation process this supplies ferrous 
iron to the leaching solution, which is necessary because there is not 
sufficient iron dissolved from the ore. This becomes oxidized to ferric 
iron in the tank house and thus provides the solvent for the sulfide 
copper. 

The only solution discarded is the moisture in the tailings, and 
therefore the wash water and wash-water advances are carefully 
balanced (Table 12). Usually a batch wash is used, i.e., the wash 
liquor is added, circulated, and then drained before the next wash is 
added. 

TABLE 12 
VOLUME BALANCE (1931), INSPIRATION 





Gallons per 
Ton of Ore 




Gallons per 
Ton of Ore 


Entering moisture (in ore) 
Fresh water added in 
wash 
Total incoming water 


16 49 

15 87 
32 36 


Moisture in tailing 
Evaporation per ton per 
day 
Total water lost 


24.73 

7.63 
32.36 



The ferric sulfate content in the leach solution is held at 7.5 grams 
per liter with 10.0 as a maximum and 5.0 as a minimum. This is 
necessary to insure good leaching of the chalcocite. New acid is re- 
quired in the amount of about 23 pounds of 60 Baume acid per ton 
of ore leached. 

During the four summer months the leach solution is not heated 
and the temperature averages 38 C. From October to May the solu- 
tion is heated to about an average of 35 C (42 C on the oldest 
charge). This is important in the leaching o' the sulfides, for with a 
cold solution (20 to 22 C) only 40 to 50 per cent of the sulfides 
dissolve, but at 35 C, the extraction rises to 75 per cent, other things 



360 HYDROMETALLURGY 

remaining equal. Metallurgical results for 1930 and 1931 are shown 
in Table 13. 

TABLE 13 
METALLURGICAL RESULTS OF SAND LEACHING AT INSPIRATION 

(Per Cent) 

Feed: 1930 1931 

Oxide copper 658 625 

Sulfide copper 0550 0.681 

Total copper 1.208 1.306 

Tailing: 

Oxide copper 020 020 
Water-soluble 

copper Trace Trace 

Sulfide copper 0.092 0142 

Total copper 112 0.162 

Extraction * 

Oxide copper 9696 96.80 

Sulfide copper 83 27 79 15 

Total copper 90 73 87 60 

Electrolysis. The commercial division contains 120 electrolytic 
tanks each 33 feet long, 4 feet wide, and 4 feet 3 inches deep. These 
are divided into 8 banks of 15 tanks each, the electrolyte flowing in 
series through the banks. Each bank has a solution circulation of 
about 1200 gallons per minute or 80 gallons per tank per minute. 
Each tank has 95 cathodes and 96 anodes spaced 4 inches from center 
to center of cathodes; this makes a total of 11,400 commercial cathodes 
and 11,520 anodes. 

The anodes are made of a lead-antimony alloy containing 8 per 
cent antimony; the submerged section is 38 by 40 inches and % inch 
thick. 

The stripper tanks constitute, in reality, a small refinery of 20 tanks. 
Each tank has a capacity of 95 cathode blanks and 96 anodes, making 
a total of 1900 blanks, capable of producing 3800 starting sheets per day. 

Soluble anodes are used in the stripper tanks. These are cast from 
blister copper at the International smelter. Usually a special electro- 
lyte is used for making starting sheets, because the commercial electro- 
lyte is not suited for the purpose. Blanks are stripped every 24 hours ; 
they weigh 11 to 12 pounds before the loops are attached. The 
cathode life varies with the current density but is about 5 days for 
maximum production; finished cathodes weigh 90 to 100 pounds each. 

The entire leaching process is built around the oxidation of ferrous 
sulfate to ferric sulfate by oxidation at the anode. The resulting ferric 



INSPIRATION 



361 



sulfate is the leaching agent for the sulfides in the ore. The " current 
efficiency " is comparatively low because of the corrosive action of 
ferric ions on the cathode deposit. Actually, the electrolytic cells are 
considered to have a double function reduction of copper at the 
cathode and oxidation of ferrous ions at the anode and instead of 
" current efficiency " as we have used it before, two values are reported, 
namely cathode efficiency and anode efficiency. The cathode efficiency 
is the ratio (in per cent) of the actual copper deposit to the theoretical 
deposit and is the same as " current efficiency " ;i,s reported at other 
plants. The anode efficiency is the ratio of the weight of ferrous 
sulfate oxidized to the weight of ferrou* sulfate which could theoretically 
be oxidized (also expressed as per cent). The a lode efficiency depends 
upon the condition of the anode surface, which must be cleaned at in- 
tervals to maintain the anode efficiency. Table 14 gives some of the 
tank house data for the years 1927 to 1931. 

TABLE 14 
TANK HOUSE RESULTS AT INSPIRATION 





1927 


1928 


1929 


1930 


1931 


Current density, amp/sq ft cathode. 


14 2 


11 8 


14 6 


13.3 


11.9 


kw-hr per pound of copper. 


1.641 


1 540 


1.695 


1.615 


1.431 


Cathode efficiency, per cent. 


61.4 


64 7 


61.8 


64.1 


67.9 


Anode efficiency, per cent. 


41.1 


43.3 


60.7 


52.0 


54.0 


Electrolyte, in: 












Copper, grams/liter. 


28 2 


23 4 


28 2 


33.1 


26.4 


Total acid, grams/liter. 


41.4 


30.6 


25.5 


26.9 


33.8 


Ferric iron, grams/liter. 


5.9 


5.3 


4.5 


3.1 


4 


Electrolyte, out: 












Copper, grams/liter. 


22 3 


18 2 


23 1 


26 6 


21 6 


Total acid, grams/liter. 


56 6 


45 


42 4 


38 4 


41 1 


Ferric iron, grams/liter. 


10 2 


9 3 


10 


9.1 


8 6 


Total iron in electrolyte, grams/liter. 


17.3 


16 


18 6 


16 4 


20.0 



Cementation Launders. There are nine double-section cementation 
launders 60 feet long, 20 feet wide, and 5 feet deep charged with baled 
tin cans which serve as a precipitant. Wash waters from the leaching 
plant are treated here to precipitate the copper, and the tailing water 
is returned to the washing and leaching circuit. The function of the 
cementation plant is twofold: (1) to maintain the desired concentra- 
tion of iron in the leaching solutions, and (2) to provide copper-free 
solution for washing the ore so that only the minimum amount of 
fresh water need be added to the system. No solution is discarded 
except that entrained in the tailings. 



362 



HYDROMETALLURGY 



The amount of cement copper produced is about equal to the weight 
of blister copper anodes used for making the starting sheets and there- 
fore amounts practically to an exchange of copper with the smelter. 
The amount of scrap iron consumed is 1.9 to 2.2 pounds per pound of 
cement copper; this high figure is due to the free acid and ferric sulfate 
in the wash solutions. 



Concentrates ^ Water to Cone. Dept v 


| to Smef 


ter f 
tlon Tailings | Add Acid 


Classifier * Flota 




* 'i i s*^ 


V/ x> <^p* >r Concentrator ^ T ^ I"!* )l \ 

Bowl Classifier Thickiner "*-" +-^4+ /l--^ O""" 1 
Slime Pump | I an ^? j Agitators f 

R i 
Thlckeners-Counter-current Washing ^ ,JL^ ^^"CIKY" 


R +-* 

r 4-C3f- u * 

Ti T-*mr! }^ 


c i u., ., t ' ,i-~'*! T IT^p ' 


L-as i^j^pjL-W- 1 ?^ 

^ Vpi ^ '-^^ M> ^ ^.^j 

._^_H J- 1 

ron launders I 


Tailings '"-y-^ 1 Ip! 1 4 


*- -^- -^ tyj\. 

Make-up Water T 


-, ^ -j-jj 


i i j 


Legend U ..-< 
Flow of Pulps P 
Flow of Solutions 
R Repulpers 
P Pumps 


^ Cement Copper to Smelter 





(Aldrich and Scott, Am. Inst. Afin. and Met. Eng. Trans , Vol. 106, p 666, 1933) 
FIG. 16 Flowsheet of the Slime Treatment Plant, Inspiration. 

Slime-Treatment Division. The slimes from the washing plant are 
treated first by flotation to remove the sulfides, and the flotation tail- 
ings are then treated by agitation in acid solution to dissolve the oxide 
copper. The leached pulp is washed by countercurrent decantation in 
four thickeners, and the copper in solution is precipitated on scrap iron. 

Figure 16 is a flowsheet of the slime-treatment division. The agita- 
tors consist of circular lead-lined wooden tanks, and the agitator 
blades are of rubber-covered steel. Four 150-foot Dorr traction thick- 
eners are used for washing, the overflow from the first thickener con- 
tains 2.73 grams of copper per liter and goes to precipitation, the spigot 
product from the fourth thickener goes to waste, and the entrained 
solution carries 0.21 gram of copper per liter. Table 15 gives the 
combined recoveries for the entire plant both the leaching division 
and slimes-treatment plant. 



KATANGA 363 

TABLE 15 
COMBINED RESULTS FROM LEACHING AND SLIME PLANT AT INSPIRATION 

Original feed to leaching plant: 1980 1931 

Oxide copper . .... . per cent 0.704 0.655 

Sulfide copper .... do . 0.536 0.665 

Total copper .. . do . 1.240 1.320 

Combined tailing from leaching plant and slime plant: 

Oxide copper . . per cent 0.022 0.025 

Sulfide copper do 0.093 0. 140 

Total copper do . . 0.115 0.165 

Combined extraction : 

Oxide copper . . . ..percent 96.875 96.183 

Sulfide copper ... ... do . 82 649 78.947 

Total copper . . do . 90 726 87 500 

Copper recovered, Ib/ton of ore ... . . 22 . 50 23 . 10 

Katanga. 16 In our previous discussion of the metallurgy employed 
in the Province of Katanga, Belgian Congo, we mentioned the fact 
that leaching had been employed to treat some of the oxide ores of the 
Katanga district. The Panda leaching plant, which has a capacity 
of 30,000 metric tons of copper per year, differs from the plants we 
have considered so far in two important respects. 

1. The leaching is carried on entirely by agitation. The ore con- 
tains fine slimes and large quantities of malachite, which effervesces 
strongly with acid. These two constituents would make it difficult to 
employ ordinary percolation leaching, so it was decided to use an all- 
agitation leach. Although the ore is ground finer than is customary 
at most leaching plants, we can hardly refer to this as an " all-slime " 
process, because there is plenty of coarse material in the ore as it is 
leached. In fact this coarse material made it necessary to modify the 
regular Pachuca tank to prevent segregation during agitation. 

2. The ore is of a much higher grade than ores commonly treated 
by leaching 6.5 to 7.0 per cent copper as against 0.9 to 1.5 per cent 
at other leaching plants. 

The Ore. The ore is almost completely oxidized, with malachite as 
the predominating copper mineral. Minor amounts of azurite, chryso- 
colla, cuprite, and native copper are in evidence, together with traces 
of sulfides. The gangue is siliceous in character and consists of shales, 
sandstones, and quartzose rock. 

16 Wheeler, A. E., and Eagle, H. Y., Development of the Leaching Operations 
of the Union Mmiere du Haut Katanga: Am. Inst. Mm. & Met. Eng. Trans., Vol. 
106, p. 609, 1933. 



OrelnB.R. Car / 



fl'x 36"jaw Crushers (Set 4? g\ 



Underslze ( 
2-Hoppers 80 M T Cap. (Ea.) 
2.48"Belt Feeders Q 



1 f \ 

Underti/e ( 1+ O') Oversize ( V'-fT) Oversize 


-Art 


Onderslze ( 1 


1 1 





I 


. t 


J 



'> 



1-24 Belt Conveyor/ 



Secondary Crushing Plant 



2-No. 5 Gyratorys (Set 2 
2-60"x 12io"rrom(net 



2-Sets 54x 16 Rolls (Set l' 

c 







1-24'Belt Conveyor /fo 

~~~ 



1.24 Belt Conveyor 







Ore Storage Plant 

H ^~^ "" 

.3-24 Distributing Conveyors (1 In Uie) 
3-Ore Piles 3800 M.T Capacity (Ea.) 



2-Recliimlng M< 



lachines 



Q In 0e) 



T 







3-24 Reclaiming Conveyors (1 In Use) / 

; } 

1-24 Collecting Conveyor /^\ 
lelt Conveyor sft. 



1 Gfifl 



Drying and Grinding Plant 

5-Hoppera 30 M T. Capacity (Etch) 

5-24 ;/ Blt Feeden ( 
WoodFuct 



(4 In Use) ( 




Use) , 





j 


S> 


^ 

56 i 


S*14"x /'Bucket Elevators (4 In Use) 


5-16"Screw Feeders (4 In Use) (. 


10-8^0"Hummer Screens (8 In Use) 


^ Oversize (-1+20 Mesh) ^ Underslze ( 


| 
-20Mesh-fO) /-^. 


V | W 
l2-0"Rod Mills (4 in Use) ^-^ 


2000 M.T, Capacity 




| VB/ 
t Fine Ore Bins 


* 



To Leaching Plint 
(Wheeler and Boole, Am. Inst. Min. and Met. Eno. Trans., Vol. 106, p. 685, 1938} 

FIG. 17-a. Flowsheet of Crushing and Grinding Plant, Panda. 

364 



Leachlno Plant 


Leaching Section 


From Electrolytic Plant 


Purification Section t 


l.ia'Wt Feeder fij\ 5-18 Belt Feeders /TjN 


2'11-o'bla. x 35'll"pachuca Agftators /^v 5-lltf'Dla. x 27-Ji Pachuca Agitators /jj\ 


2-11-O'Dla. x 35'n "Pachuca Agitators fi\ 5-1l'-0"oia. x 27-5' Pachuca Agitators ^ 


2-11 Voia. x 35'll" Pachuca Agitators ^ Ml'-GDIa. x 27-V Pachuca Agitators /^ 


2.l1-o"oi8. x 35-1l" Pachuca Agitators ^ 5-lltf'oia. x 27-5 Pachuca Agitators ^ 


' I 
1-6 Duplex Dorr Classifier (Tt\ 


\ 
s (- 


I 




Slimes (-vOO Mesh)/^ Sands (-f 200 Mesh) /^ 


6.4'6"0orr Bowl Classifiers fi\ 




V v JSIIm 


-200 Mesh) /^ Sands (-1-200 Mesh) Wssh H a O <fa 


1 70 Dla Dorr Decanting Thickener ^\ 


f 


\ty r~^ (23) I v*i/ 
8-4'-6"TTi'*e Deck Dorr Washing Classifiers fi\ 


Overflow ffih Underflow /Ok 


Sands (4- 200 Mesh) /^\ 


1 ^ 
Wash Solution /O\ 


^ f > ~* |.4 Duplex Oorrco Pump ^ 


v f ^ 


f^ y ,...* , J 


T 

5-70 Dla Dorr Decanting Thickeners /Jih 
Overflow/Ok Underflow /Ok 
c > 7 M^Duplex Dorrco Pump 
v ^ ^ J 


1 70'Dli Dorr Washing Thickener ^n\ 


Overflow^ UnJerflow^v 


Ldl'PUP'* 11 'orrco Pump 


j 


5-70'Dli Dorr Washing Thickeners f3j) 




1 70 ' Ola Dorr Washing Thickener ^ 


Overflow /^5\ Underflow /TTV 
M"Duplex Dorrco Pump 


J 1 
Overflow /f7\ Underflow (52) 

^ ^ Wash HjO /*"\ 1-4 Duplex Dorrco Pump 




, j 


1 ^''^ ^ 


5-70'Dli Dorr Waihing Thickeners ^\ 


1-7o'ori Dorr Washing Thickener f^\ 


Overflow ^ Underflow ^\ 
* ' Wash H,0 ^s. 1.4 v Duplex Dorrco Pump 


Overflow ^\ Underflow Q 


^ ^ (l-Vbuplex lorrco Pump 


t ^ 


} 


^ Sluicing H|0 /^\ 
[ 


5-70'Dia Dorr Washing Thickeners ^ 




1 
Overflow^ 


1-4 Duplex Dorrco Pump 


J 


\ 

Tailings Launder /gj\ 

Tailings Pond 1-7o'oia Dorr Clari 


yng Thickener ^ 




Overflow /O 
t 


I 
j 

2-4" 


^rTp 


2'70'Dla. Storage Tanks /O 


t 


l ^JUJ 

Solution /ov 


luplex Dorrco Pumps 




| 1 
3-Purlflcstlon Pumps (1 tn Use) 3-Clrculallon Pumps (1 In Use) O\ 


J @) ^ 




{ 









To Electrolytic PUnt 



FIG. 17-6. Flowsheet of Leaching Plan f , Panda. 



365 



366 



HYDROMETALLURGY 



The ore is crushed and ground to approximately 20 mesh before it 
goes to the leaching agitators. The crushing plant contains one pri- 
mary jaw crusher, two No. 5 gyratories, and two sets of 54- by 16-inch 
rolls for secondary crushing. The roll product is dried and ground dry 
in 6- by 12-foot rod mills in closed circuit with a 20-mesh screen. 
Screen undersize passes to the leaching plant. 



Lufira Project 



Panda Centrale 



Transmission Line 



Transmission Line 



Substation 



Power 


Comp. Air 




To Various Oepts, 


Electrolytic Plant 


1 
M0i<g 


>t 16-Startlng Sh 




eet Tanks 


^ i 
Starting Sheets /rjv Sc 


I 

tp y^v Solution /gy\ 
2 Pumps (1 In Use) 


, - ^ . , 

144-Cathode Tanks 


\ 

J 


1 

Cathode ^\ 


jChap-Kg) 
n Use) ^ 


Furnace Refinery 






Coal Fuel ^ _Polet /ga 
2-13'x 27-V'Refinlng Furnaces (1 


Smelter 
2 


i ^^ 

\ Refined Copper /g>^ 


Copper Casting Wheefs (1 


,., 


Rejects ^ 


J^ Wire Bar Ingots, Etc. 


^T 


I 

Market 





FIG. 17-c. Flowsheet of Electrolytic Plant, Panda. 

Leaching Division. The leaching method employed at Katanga dif- 
fers in several respects from percolation leaching. Some of the im- 
portant differences are as follows: 

1. Ore and leaching solution are agitated long enough to allow dis- 
solution of the copper minerals. 

2. Washing and separation of liquid and solids are carried out in 



KATANGA 367 

thickeners; simple draining such as is used in vat leaching is not ap- 
plicable. 

3. The process is continuous with respect to both ore and solution, 
and employs the countercurrent system of leaching and washing. 

4. Purification of the solution is conducted in the leaching system 
by means of an " acid leach " and a " neutral leach." 

Figure 17 shows the flowsheet of the leaching plant, and Figure 18 
shows the general plant layout. The plant contains 28 agitators of 
the standard cone-bottomed Pachuca type built of steel and lined with 
10-pound chemical lead sheet. They are equipped with a special dis- 
charge device to prevent coarse material from building up within the 
agitator and are provided with a supply of air ;tt 30-pound pressure. 
Twenty of the agitators are arranged in five parallel sections each with 
four agitators in series; these constitute the five acid leach sections. 
The other eight agitators make up the purification section, which con- 
sists of two parallel groups of four tanks each arranged in series. 

The bulk of the ore coming from storage is distributed by means of 
five conveyor belts to the five parallel groups of acid Pachucas. The 
spent electrolyte flows to the agitator buildings and is distributed to the 
five acid sections by means of a five-compartment weir box. The solu- 
tion feed to each section joins the ore feed to that section in a mixing box 
immediately ahead of the first agitator in the series. The pulp then 
passes through the four agitators in series, and most of the copper goes 
into solution here. 

After leaving the agitators the pulp goes to one of six 4-foot 6-inch 
Dorr bowl classifiers which classify the pulp into sands ( 4- 200 mesh) 
and slimes (200 mesh) which are washed separately. The sands are 
washed in a three-deck washing classifier which is built integral with 
the bowl classifier. There are six of these bowl classifier-washing classi- 
fier units, one for each acid section and a sixth for a spare. Fresh 
wash water is added on the third deck of the washing classifier in 
amount equal to the entrained water carried out in the sand tailings. 
The washed sand tailings from the third deck of the washing classifier 
passes directly to the tailing dump. 

The slimes from the bowl classifier are washed by countercurrent 
decantation in a series of four 70-foot Dorr thickeners. The overflow 
from the first thickener is the strong solution which passes through a 
clarifying thickener and on to the electrolytic tank house. Wash water 
is added to the last thickener, and the spigot product of the last thick- 
ener goes to the tailing dump. There are five sets of these thickeners, 
one for each acid circuit. Thus the acid circuit leaches the ore and 
washes the solution free from the solids. Spent electrolyte enters the 



368 HYDROMETALLURGY 

first agitator, and strong solution overflows the first slime washing 
thickener. The only other solution leaving the system is the water 
entrained in sand and slime tails. Wash water to replace this is added 
on the third deck of the sand-washing classifier and to the last of the 
slime- washing thickeners. 

The only impurity in the ore which builds up enough to interfere with 
electrolysis is iron; some A1 2 3 is dissolved also, but this is removed 
with the iron. The solution is purified to keep the iron content below 
about 5 grams per liter. To attain this result a certain amount of 
strong solution is circulated through the purification section. 

In general the purification section resembles two sections of the acid 
leach section as far as equipment goes. The method of purification 
employed is to use an excess of ore to neutralize the free acid in the 
solution; aluminum and ferric salts hydrolyzc in this neutral solution 
and precipitate as hydroxides In the presence of a large excess of ore 
or concentrate the sulfuric acid is soon used up and the iron and 
aluminum salts hydrolyze. 

A1 2 (S0 4 ) 3 + 3H 2 O -> 3A1(OH) 3 + 3H 2 S0 4 
Fe 2 (SO 4 ) 3 + 3H 2 O -> 3Fe(OH) 3 + 3H 2 S0 4 

The acid formed is immediately consumed by the ore or concentrate, 
and more iron and aluminum are precipitated. Either ore or high- 
grade oxide concentrate is used for purification; the solution flowing 
through the system has all the free acid neutralized, drops most of its 
iron and aluminum, and becomes enriched in copper dissolved from the 
neutralizing agent. 

This method of purification requires an excess of ore, which means 
that much undissolved copper is left in the " tailing." Also if the solids 
were sent back to the acid leach, the precipitated impurities would 
redissolve. The bulk of the undissolved copper is in the " sands " 
(4-200 mesh), and most of the precipitated impurities are in the 
" slimes " (-200 mesh) ; so the overflow from the last of the neutral 
leach Pachucas is sent to a 6-foot Dorr duplex classifier which separates 
it into sands and slimes. The sands are sent to the head of the acid 
leach to join the new ore, the slimes are washed through four Dorr thick- 
eners in a neutral circuit, and the spigot product from the last thickener 
goes to the tailing dump. 

The leach solutions are so corrosive that the only materials which 
can be used in contact with them are lead, Duriron, rubber, asphalt 
mastic, glass, porcelain, and certain of the chrome-nickel-iron alloys. 
Thickener tanks have concrete bottoms and steel sides and are lined 



KATANGA 



369 




370 HYDROMETALLURGY 

with lead. Blades in the rake classifiers are made of Duriron, and any 
steel which could come in contact with the solution is covered with 
sheet lead. 

Electrolysis. The tank house contains 160 electrolytic tanks each 
62 feet 6 inches long, 3 feet 2 inches wide, and 4 feet 2 l / 2 inches deep, 
inside dimensions these are exceptionally long tanks. Tanks are of 
reinforced concrete lined with 2 inches of asphalt mastic. They are 
equipped with a feed pipe at one end, an overflow dam at the other, and 
a hard lead plug and seat in the bottom for draining and cleaning out. 
Conductor bars 2 by 6% inches in cross-section are supported on wood 
insulating strips on the tank walls. 

Anodes are of antimonial lead containing about 6 per cent antimony, 
and starting sheets are of the regular type made in stripper tanks using 
soluble anodes. The electrodes are arranged in three separate groups 
in each tank, the groups being in series and the electrodes in multiple. 
The entire tank house is laid out in two main electrical circuits, each 
taking normally 8000 amperes at 460 volts. Cathodes have a life of 
10 to 14 days. 

Sixteen of the electrolytic tanks are equipped so that they can be fed 
with a separate circulation of pure solution, and enough of these tanks 
are so used to produce the necessary starting sheets from soluble 
anodes. 

Refinery. The furnace refinery is provided with two 130-ton re- 
fining furnaces each equipped with a 38-foot Clark casting wheel. The 
soluble anodes used in the stripper tanks are made in the furnace 
refinery. 

Many of the significant data pertaining to the leaching and electroly- 
sis are given in Tables 16 and 17. Note the effect of the purification 
cycle on the solution, also the heavy " cut " made by electrolysis re- 
ducing the dissolved copper from 30.5 to 16.3 grams per liter. 



TABLE 16 

SUMMARY OF PLANT OPERATIONS FOR DECEMBER 1929 AT 
PANDA LEACHING PLANT, KATANGA 

Crushing and Drying 

Wet ore crushed 40,056 metric tons 

Moisture in wet ore 13 . 20% 

Cu in dry ore 6 . 57% 

Wet ore to driers 39,984 metric tons 

Moisture in dried ore . 24 % 

Leaching Division 

Dry ore fed to leaching, weight 32,631 metric tons 

Dry ore fed to leaching, copper assay 6.539% 

Dry concentrates fed to purification, weight 1,058 metric tons 

Dry concentrates fed to purification, copper assay 28.70% 
Classifier sands: 

Copper assay (dry) 0.442% 

Entrained solution 29 . 19% 

Cu in entrained solution 11 .84 g/1 

Free H^SCU in entrained solution 0.00 g/1 
Acid slime tails: 

Copper assay (dry) 0. 177% 

Entrained solution 49 39% 

Cu in entrained solution 9 . 52 g/1 

Free H2SO4 in entrained solution 0.00 g/1 
Purification slime tails: 

Copper assay (dry ) . 426 % 

Entrained solution 57 . 20% 

Cu in entrained solution 11 33 g/1 

Free H2SO4 in entrained solution 0.00 g/1 
Tailing: 

Sand tails, % of total tailing 40.54% 

Acid slime tails, % of total tailing 56.08% 

Purification slime tails, % of total tailing 3 38% 

Total weight of tails per metric ton of ore 869 metric ton 

Total weight of tails per metric ton of concentrates 491 metric ton 
Total extraction (% of total copper dissolved; 100% chemical loss) 96.53% 
Total recovery (% of total copper delivered to electrolysis; 

100% chemical loss mechanical loss) 88.34% 

Electrolytic Division 

Copper produced per kw-hr 0.471 kg (1 .03 Ib) 

Ampere efficiency 78 . 53 % 

Volts per circuit 444.69 volts 
Voltage across adjacent electrodes: 

Commercial tanks 1 . 97 volts 

Stripper tanks 0.39 volt 

Tanks in service, total 160 

Commercial 148 

Stripper 12 

371 



372 



HYDROMETALLURGY 



TABLE 17 
AVERAGE SOLUTION ASSAYS, IN GRAMS PER LITER, AT PANDA LEACHING PLANT 





Tank House 


Acid 


Purification 




Feed 


Discharge 


Agitation 


Feed 


Discharge 


Cu 


30 51 


16 26 


30 53 


30.17 


44 98 


Free H 2 SO 4 


5 50 


34.35 


12 22 


11 17 


0.00 


Fe, total 


5 63 


5 63 


6 65 


5.81 


1.99 


Fe, ferric 


5.43 


3.97 


6 34 


4.98 


1.42 


Fe, ferrous 


20 


1.66 


31 


83 


0.57 


A1 2 3 


11.00 


11 00 


11 70 


10 88 


7.50 



OTHER IMPURITIES ACCUMULATED IN SOLUTION 



Co 


Mn 


MgO 


Si0 2 


P 2 5 


CaO 


10.03 


6.50 


13 60 


1.72 


3.80 


0.34 



GRADE OF COPPER PRODUCED BY ELECTROWINNING 

The cathode copper produced by leaching and electrolysis has about 
the same purity as electrolytically refined copper. Several typical 
analyses are given in Table 18, and for the sake of comparison some 
analyses of electrolytically refined copper are included. 



GRADE OF COPPER PRODUCED BY ELECTROWINNING 373 



TABLE 18 

ANALYSES OF ELECTROLYTIC COPPEE 
ELECTROWINNING 



Sourer 


Cu 


Au 


Ag 


S 


As 


Sb 


Andes; wirebars 6 
Inspiration ; cathodes 
Chuquicamata; cathodes'* 
Chuquicamata; wirebars d 


99 9610 
99 87 
99 900 
99 9600 


Trace 


019 


0017 
045 
03 
0016 


0005 
Trace 
0002 
0002 


0002 
0.0005 
0.0004 
0.0004 





ELECTROREFINING 



Montreal East; wirebarV 


99 965 


015 


2. 


0013 


0.0001 


0001 


Ontario; wirebars-^ 


99 960 




0012 


0018 


0005 


0.0010 


Raritan; wirebars* 7 


99 95 


003 


0010 


0020 


0.0015 


0015 


Great Falls; wirebars^ 


99 955 




0034 


0019 


0015 


0009 



ELECTROWINNING 



Source 


Fe 


Co-fNi 


<>2 


Se+Fe 


Zn 


Cl 


Pb 


Sn 


Andes; wirebars 
Inspiration ; 
cathodes 6 
Chuquicamata ; 
cathodes 6 
Chuquicamata; 
wirebars c 


0013 
0.021 
0.0050 
0016 


0001 
0006 


0351 




Trace 
0003 
Ti ace 










Trace 
005 










0.0016 
0008 


0011 
0006 




0320 





ELECTROREFINING 



Montreal East; 
wirebars 


0.0016 


0.0002 


035 


0002 






0.0007 




Ontario ; wirebars 6 
Raritan ; wirebars^ 


0013 
0.0025 


0016 
0015 


0300 
035 


0032 






0.0003 





Great Falls; 
wirebars^ 


0026 


0014 


0.030 































Au and Ag in ounces per ton , all others in percentage 

6 Callaway, L A , and Koepel, F. N , Metallurgical Plant of the Andes Copper Mining Co.: Am. 
Inst Mm and Met Eng Trans., Vol. 106, p 726, 1933 

c Aldrich, H. W , and Scott, W. A., The Inspiration Leaching Plant: Am. Inst. Mm. and Met 
Eng. Trans , Vol. 106, p 650, 1933. 

d Campbell, T C , A Brief Description of the Reduction Plant of the Chile Exploration Company at 
Chuquicamata, Chile, S. A : Am Inst Mm and Met. Eng Trans , Vol 106, p. 559, 1933. 

* McKnight, H. S., Montreal East Plant of Canadian Copper Refiners, Ltd Am. Inst. Min and 
Met. Eng Trans , Vol. 106, p 352, 1933. 

t Benard, Frederic, Electrolytic Copper Refinery of Ontario Refining Company, Ltd Am Inst 
Mm. and Met. Eng. Trans., Vol. 106, p. 369, 1933. 

Burns, W. T , Refining Anaconda Copper at Raritan and Great Falls' Eng. and Min. Jour , Vol. 
128, No. 8, p. 306, 1929. 



374 HYDROMETALLURGY 

OTHER LEACHING METHODS 

There have been many leaching operations utilized at one time or 
another in the metallurgy of copper, but we have confined our discus- 
sion principally to methods which are being used commercially to ex- 
tract copper from its ores. To present commercial practice, the fol- 
lowing statements apply: 

1. The most common type of leaching is sulfuric acid leaching used 
on low-grade ores in large-scale operations. 

2. Certain special ores (or tailings) containing either native copper 
or copper carbonates are treated by ammonia leaching. 

3. Sulfides are not attacked by sulfuric acid alone but require an 
oxidizing agent in addition. The oxidation may be carried on as a 
separate operation (weathering, roasting) or the oxidizing agent may 
be dissolved in the leach solution (ferric salts). 

4. Electrolytic precipitation is used with sulfuric acid leaching. 
The low-grade solutions from heap leaching and mine waters and the 
pregnant solution from an ammonia leach are not adapted for electro- 
lytic precipitation. 

5. Acid leaching has been applied to ores containing both oxides 
and sulfides (e.g., at Inspiration) , but all-sulfide ores are still treated 
by milling followed by smelting of the concentrate. 

We have akeady mentioned the question of using leaching methods 
for treating high-grade sulfide concentrates, and we shall conclude this 
chapter with a brief description of an experimental plant designed for 
this purpose. Up to the present, no full-scale commercial plant has 
been put into operation to treat copper sulfide concentrates by leaching. 

Bagdad, Copper Corporation. 17 The pilot plant at the Bagdad 
properties near Hillside, Arizona, treats two types of concentrates (1) 
a chalcopyrite concentrate containing 25 per cent copper and (2) a 
chalcocite concentrate assaying as high as 45 per cent copper. These 
concentrates are first given a special roast which aims to put the cop- 
per in the form of CuO and the iron in the form of Fe 2 3 . Under 
these conditions the copper is soluble in sulfuric acid and the iron only 
slightly soluble. 

Roasting, First Stage. The first stage is a low-temperature roast 
designed to remove all the sulfide sulfur and convert copper and iron 
to oxides and sulfates. The first oxidation of sulfur begins at 500 F 
(260 C) with the expulsion and burning of the free-atom sulfur in 
pyrite and chalcopyrite. Oxidation continues to the end of this stage, 

17 Baroch, C. T., Hydrometallurgy of Copper at the Bagdad Property: Am. 
Electrochem. Soc. Trans., Vol. 57, p. 205, 1930. From a paper presented before the 
1930 meeting of the American Electrochemical Society, Inc. 



BAGDAD COPPER CORPORATION 375 

at which point the roast is " dead/' i.e., it shows no glowing particles 
or sparks when rabbled, and there are only traces of S0 2 in the furnace 
gases. The calcine consists of cuprous oxide, Cu 2 0; ferrous oxide, 
FeO; a magnetic oxide, probably FcOFe 3 4 or FeOFe 2 3 ; and some 
cupric and ferrous sulfates. The calcine at the end of the first stage 
is black or gray, sometimes with a brownish tinge, and is so magnetic 
that almost the entire mass can be picked up with a magnet. 

Temperature control determines the amount oi sulfates formed. 
From 700 to 720 F (370 to 380 C) appears to oe the temperature 
of maximum sulfate formation, and as much as 60 per cent of the cop- 
per may be present as sulfate at 700 F, but at 850 F (455 C) only 15 
per cent of the copper will be present as the suifate. 

Although the calcine is strongly magnetic, it appears that the iron 
is not present as a true magnetite as long as the temperature remains 
below 850 F. Above this temperature a true magnetite forms, prob- 
ably by the reaction: 

FeS + 10Fe 2 O 3 -> 7Fe 3 4 + SO 2 

This magnetite is difficult to reoxidize to Fe 2 3 . 

The object of the first roasting stage, therefore, is to keep the tem- 
perature below 850 F until all the sulfides are decomposed; after the 
sulfides are gone there is no danger of reducing Fe 2 O 3 to refractory 
magnetite. 

Roasting, Second Stage. The second stage is an oxidizing and de- 
magnetizing period, and its object is to produce the higher oxides of 
copper and iron. 

2Cu 2 O + O 2 -> 4CuO 

4FeO + O 2 - 2Fe 2 3 
4FeO-Fe 2 O 3 + () 2 -> 6Fe 2 3 

True crystalline magnetite, if present, does not readily oxidize to 
Fe 2 3 . Moreover, if sulfides and the concomitant SO 2 are present, 
some of the cupric oxide will be reduced back to Cu 2 0, thus: 

2CuO + SO 2 -> Cu 2 O + S0 3 

The presence of Cu 2 in the finished calcine is undesirable for leaching, 
as we shall see later. 

Temperatures during the second stage are not critical, and the oxida- 
tion depends principally upon the amount of ox>gen in the roaster at- 
mosphere and the use of thorough rabbling to insure contact of the 
oxygen with the charge. The best temperature for the second roast is 



376 HYDROMETALLURGY 

about 980 F (530 C), but the temperature may range from 850 to 
1000 F (455 to 540 C) . 

As the calcine progresses through the second stage the magnetic 
properties become less pronounced and finally disappear entirely, and 
the color gradually changes to the brilliant red of ferric oxide; in 
fact the color of the calcine can almost be used instead of control 
assays; the color comparisons must be made on cooled calcine, as all hot 
calcines are dull black. 

Roasting, Third Stage. The third stage is essentially an extension 
of the second stage in which the temperature is raised to about 1040 F 
(560 C), and its purpose is to decompose the water-soluble sulfates 
of iron. Anhydrous ferric sulfate begins to decompose into ferric 
oxide and S0 3 at about 300 F and cannot exist above 716 F (380 C) 
as normal ferric sulfate. However, it forms a basic sulfate, probably 
Fe 2 3 -2S03, which does not decompose completely below 1040 F 
(560 C) . Much of the liberated S0 3 combines with CuO to form 
CuS0 4 . A temperature of 1200 F (650 C) is required to decompose 
CuS0 4 , and it is found that heating to this temperature invariably 
forms a black magnetic iron oxide which is fairly soluble in acid 
solutions. 

The control of the entire process depends upon the roasting opera- 
tion, and the most critical point in the roast is the transition point 
between the first and second stages. If the temperature is allowed to 
rise above 850 F before all the sulfide sulfur is removed the calcine 
will invariably show low copper solubility and high iron solubility. 
The principles involved in this roast are covered in U. S. Patent 
No. 1,674,491, issued to Herbert E. Wetherbee. 

Table 19 gives the analyses of the calcines produced from the 
chalcocite and chalcopyrite concentrates. 

The results given in Table 19 are from laboratory data, and al- 
though larger-scale tests were not quite as good, the results of all 
the tests seemed to indicate that extractions of 97 per cent could 
be made on a regular commercial scale. 

Leaching and Electrolysis. The leaching and electrolysis of properly 
roasted calcine do not appear to present any difficulties over present 
commercial leaching practice. It is indicated that these high-grade 
concentrates can best be leached by agitation followed by thickening 
and filtration of the tailing. It should be noted that the tailing 
produced would amount to only about 50 to 60 per cent of the calcine 
because of the large amount of soluble material in it. 

One of the principal sources of copper losses is the presence of Cu 2 
in the calcines. This compound does not represent the highest stage 



BAGDAD COPPER CORPORATION 



377 



TABLE 19 

ANALYSES OF ROASTED CHALCOCITE AND CHALCOPYRITE CONCENTRATES, 
BAGDAD, PER CENT 





Chalcocite 


Chalcopyrite 


Concentrates 


Calcines 


Concentrates 


Calcines 


Cu, total 
Cu, acid soluble 
Cu, water soluble 
Fe, total 
Fe, acid soluble 
Fe, water soluble 
S, total 
S, water soluble 
Insoluble 


37.50 


30.30 
30.20 
21 00 
19 20 
0.10 
Trace 
11 42 
11.22 
9 10 


26.10 


25.00 
24.70 
19.10 
27.20 
0.44 
Trace 
8 70 
6.36 
9.45 






23 30 

30.20 
11 00 


28.30 




33.30 


9.80 



NOTE. In the case of the chalcocite concentrate, 99.7 per cent of Cu in the calcine is acid soluble 
and 69 3 per cent is water soluble, for chalcopynte, 98 8 per cent of the Cu in the calcine is acid soluble 
and 76 4 per cent is water soluble 

of oxidation of copper, and consequently only half of it will dissolve in 
sulfuric acid. 

Cu 2 + H 2 SO 4 -> CuS0 4 + H 2 + Cu 

The precipitated copper will not be dissolved except by long con- 
tinued agitation with hot sulfuric acid or by the addition of some 
expensive or detrimental oxidizing agent (e.g., ferric sulfate). 

The large amount of water-soluble copper (CuS0 4 ) in these calcines 
indicates that electrolysis of the leach solutions will produce an excess 
of free acid; where acid-consuming oxidized ore is present, this could 
be leached with this excess acid. Electrolytic precipitation should 
work successfully and should produce copper of standard " electrolytic " 
grade. 

This process should be applicable to a wide variety of copper ores 
and concentrates, as the principal impurity, iron, is efficiently handled 
in the roasting; arsenic and antimony could also be controlled by the 
addition of a purification step following the leaching. 

Recovery of Gold and Silver. The Bagdad, ores contain only traces 
of gold and silver, so no effort was made to recover these metals. 
Laboratory experiments, however, seemed to indicate that the residual 
copper in well-washed tailing did not behave as a cyanicide, so that the 
cyanidation process might be employed where the concentrate con- 
tained gold. Chlorination methods might also be used for gold 
recovery. 



378 HYDROMETALLURGY 

For calcine tailings containing considerable silver, a chloridizing 
roast would probably be indicated, followed by a leach and precipita- 
tion on scrap iron, as in the Longmaid-Henderson process. This 
method would recover the residual copper in the tailings as well as 
the silver. 



CHAPTER X 
PROPERTIES OF COPPER* 

PHYSICAL PROPERTIES 

General. Copper and gold are the only two metals which are strongly 
colored. Copper has a reddish or rose color on a fresh surface, but old 
surfaces often have an orange tinge due to a film of cuprous oxide. 
Molten copper is light green, and very thin sheets of copper appear 
green by transmitted light because the green light passes through while 
other light rays are largely absorbed. Very finely divided copper is 
black, as are all finely divided metals. 

Crystal faces, polished surfaces, and the surface of liquid copper 
display a metallic luster, but roughened surfaces, such as cathode 
deposits, and finely divided copper powder show no luster. 

Crystallization. Copper crystallizes in the face-centered cubic 
lattice pattern, 2 - 3 and the side of the unit cube in the crystal lattice 
is 3 597 Angstrom units ; the unit cube contains four atoms. Twenty 
other elements crystallize in the face-centered cubic pattern, including 
the following metals: 

Silver y-Iron Platinum 

Aluminum Iridium Palladium 

Gold ^-Nickel Rhodium 

a-Cobalt Lead 

The face-centered cubic arrangement corresponds to one of the two 
possible patterns obtained by the closest packing of spheres (the other 
gives the "hexagonal close-packed" pattern). The metals in this 
class are usually rather ductile. The most ductile metals have the 

1 Data quoted in this chapter are from the following sources, with index num- 
bers as follows: 

2 Wyckoff, R. W. G., The Structure of Crystals: Reinhold Publishing Cor- 
poration, New York, 1931. 

3 Handbook of Chemistry and Physics, 19th ed., Chemical Rubber Publish- 
ing Co , Cleveland, 1934 

4 Metals Handbook, 1936 ed., American Society for Metals, Cleveland. 
5 Eshbach, O. W., Handbook of Engineering Fundamentals: John Wiley and 

Sons, Inc., New York, 1936. 

379 



380 PROPERTIES OF COPPER 

face-centered cubic crystal lattice (silver, copper, lead, aluminum, 
palladium, etc.). 

Copper has no allotropic modifications, as do some other metals 
(e.g., iron, cobalt, nickel, chromium), and therefore has no " critical 
temperatures " at which changes take place in the crystal pattern. 
Many heat-treating operations used on metals and alloys depend 
upon allotropic changes, but these have no application to copper. 

When copper is cold-worked it becomes harder, but the hardness 
can be removed by annealing. The softening effect is accompanied by 
the formation of a new crop of small equi-axed crystals replacing 
the elongated crystals of the cold-worked metal. The lowest tempera- 
ture at which this phenomenon takes place is the recrystallization 
temperature, and the recrystallization temperature 4 for copper usually 
ranges from 200 to 400 C. 

Highly purified copper will recrystallize at 100 C if annealed for a 
long time, but commercial copper contains impurities, notably silver, 
antimony, and arsenic, which raise the recrystallization temperature. 
Electrolytic copper begins to recrystallize at about 200 C (400 F) , 
and Lake copper or other arsenic- or silver-bearing copper recrystal- 
lizes at about 350 C (550 F). 

In spite of its high purity, the recrystallization temperature of 
oxygen-free copper is higher than that of ordinary tough-pitch electro- 
lytic copper. 

Rolled copper normally shows a small grain size (0.02 to 0.05 mm 
in diameter) and will not possess any directional properties; cold-rolled 
copper will have elongated or distorted grains, but these will be ori- 
ented at random as far as their crystal axes are concerned. 

Density. 4 The density of commercial copper samples will range 
from 8.4 to 8.94 grams per cc. Cast tough-pitch copper contains 
about 3 to 5 per cent voids or gas holes, by volume, and such copper 
will have an apparent density of 8.4 to 8.7. The voids are closed up 
during rolling, and after working and annealing the density of tough- 
pitch copper will be 8.90 to 8.93, depending on the oxygen content; 
copper containing 0.03 per cent oxygen has a maximum density of 8.92. 

Deoxidized and oxygen-free copper freezes without the voids and 
gas holes found in tough-pitch castings, and accordingly there are 
always deep pipes or shrinkage cavities in deoxidized or oxygen-free 
copper castings, caused by the volume shrinkage. The density of these 
coppers will range from 8.80 up to 8.90 for castings, and worked pieces 
will approach the maximum of 8.94. For commercial coppers contain- 
ing over 99.85 per cent copper, the following formula employed by the 

4 Metals Handbook, 1936 ed., p. 1061, American Society for Metals, Cleveland. 



MECHANICAL PROPERTIES 381 

American Brass Company gives the density at 20 C. 
d - 8.933 - 0.44 (100 - %Cu) 

Liquid copper has a density of 7.93 at the melting point, and 7.53 
at 1600 C. 

Mechanical Properties. Copper is one of the most ductile of metals 
and is also rather malleable. Pure copper is not particularly strong 
or hard in the soft or annealed condition, but both strength and hard- 
ness are increased considerably by cold working. 

Strength. 4 The tensile strength of pure copper (containing 0.015 
per cent oxygen) in the form of an annealed rod % inch in diameter 
is 31,790 pounds per square inch. This may be taken as the strength 
of " pure " copper in the annealed condition. The tensile strength of 
other specimens will depend on the chemical composition, heat treat- 
ment, and mechanical treatment. For commercial copper, the tensile 
strength of the hot rolled or annealed metal is about 30,000 to 36,000 
pounds per square inch. 

Cold working increases the tensile strength to a maximum of about 
70,000 pounds per square inch for severely cold-worked copper. 

Hardness. 5 Copper is one of the softer metals. Soft copper in the 
annealed or hot-rolled condition will have a Brinell hardness number 
of about 42. Cold-rolled copper is harder and will have a hardness of 
about 103 Brinell. 

Ductility. Annealed or hot-rolled copper is a very ductile metal 
and will show an elongation of about 50 per cent in length (2 inches) 
when pulled in a tensile machine. Cold working reduces the ductility, 
and severely cold-worked copper will show an elongation of only 5 
per cent. Figure 2 shows the effect of cold working on tensile strength 
and ductility (the latter measured by both per cent elongation and 
per cent reduction in area of the specimens). Copper is also very 
malleable but there is no quantitative criterion for malleability. 

Ductility specifically refers to the ability of metal to be drawn into 
wire, but the term is often used to mean the same as /ormabih'fy, which 
is a more general term and refers to that property of the metal which 
determines its response to all forms of mechanical working. The 
ductility is measured by the elongation and reduction of area of a 
tensile specimen as we have noted; other special tests are used to 
determine the formability for example, the " cupping test " to 
measure the amenability of sheet metal for cold-drawing or stamping. 

4 Idem, p. 1065. 

5 Eshbach, 0. W., Handbook of Engineering Fundamentals, p. 11-44: John 
Wiley and Sons, Inc., New York, 1936. 



rKU.rii.it ii.ii.e5 u* 

Oxygen-free copper is notably more ductile than tough-pitch copper 
and is used for " deep drawing/' where the metal must undergo severe 
cold working. 

Elastic Properties. Soft copper has very little elasticity or resilience, 
but both of these properties are increased if the copper is cold worked. 
Soft copper has an " elastic limit " or " yield point " of 5000 to 17,000 
pounds per square inch, depending on the convention used in defining 
these terms; there is very little true elastic deformation, and even 
small loads produce some permanent deformation. Cold-rolled copper 
will have a yield point of 44,000 to 48,000 pounds per square inch. 
Young's modulus for hard copper is about 16,000,000 pounds per 
square inch. 

Endurance Limit. 4 The endurance limit, which measures the 
maximum stress below which the metal will not fail from repeated 
stresses or " fatigue," is 10,000 pounds per square inch for soft copper. 
Cold-worked copper may have an endurance limit as high as 20,000 
pounds per square inch, but 15,000 is probably a more conservative 
figure. These values are all for commercial tough-pitch copper. 

Weldabihty, Etc. Copper can be silver soldered, brazed, or welded, 
and the welding may be done by either the arc or oxyacetylene method. 
For any welding operation where the copper is exposed to the action 
of reducing gases and high temperatures it is best to use deoxidized 
or oxygen-free copper. Even the small amount of oxygen in tough- 
pitch copper will cause embrittlement of the metal due to the segrega- 
tion of cuprous oxide at the grain boundaries and reduction of the 
oxide by the gases. 

Copper cannot be cut by the oxygen lance, as can iron and steel, 
because of its high thermal conductivity; heat is carried away so 
rapidly that it cannot be localized at the spot where the metal is to 
be cut. 

" Hardening " and " Tempering " of Copper. We have been re- 
ferring to hard and soft copper in the previous discussion, and in every 
case the term " hard " copper has meant copper hardened by cold 
working; there is no other way to harden pure copper. 

The so-called lost art of " hardening " or " tempering " copper still 
comes in for its share of attention, although it has been pretty well 
demonstrated that the hardening of copper by heat treatment is im- 
possible. The hardening of steel, for example, depends largely upon 
the fact that iron exists in several allotropic forms; copper, however, 
has no allotropic modifications. Moreover, it is doubtful if any pure 
metal can be hardened by quenching (or other heat treatment) even 

4 Metals Handbook, 1936 ed., p. 1071, American Society for Metals, Cleveland. 



THERMAL PROPERTIES 383 

though it does possess allotropic modifications. Pure iron cannot be 
hardened by quenching, although steel (an iron-carbon alloy) can be 
so hardened. 

Annealing is the only form of heat treatment applied to pure copper, 
and it is used to soften copper made hard by cold working. Certain 
copper alloys can be hardened by heat-treating methods, but these are 
fundamentally different from pure copper even though the amount of 
the alloying ingredient may be very small. 

Thermal Properties. Table 1 lists the important thermal constants 
of copper. Copper is the second best heat conductor of all the metals, 
being surpassed in this respect only by silver. Comparative values 3 
for thermal conductivity for different metals at room temperature 
are as follows: 

Heat Conductivity 
(cal/cm 2 /cm/sec/ C) 



Silver 


1.006 


Copper 
Gold 


0.918 
0.705 


Aluminum 


0.480 


Iron 


0.161 



TABLE 1 
THERMAL PROPERTIES OF COPPER* 

Melting point 1083 C (1981 .4 F) 

Boiling point 2325 C (4217 F) 

Latent heat of fusion 50 46 cal/gram (90.83 Btu/lb) 

Specific heat (25 C) a 0919 cal/gram/C 

Linear coefficient of expansion (20 C)* 16 42 X 10- 6 /C 

Thermal conductivity (20 C) c . 923 cal/cm 2 /cm/sec/C 

a A formula for specific heat which applies in the range to 100 C is 

c = 09088 (1+0 000534 It - 00000048* 2 )/cal/gram/C 

It will be noted that there are discrepancies in these different values, due principally to small vari- 
ations in composition and condition of the samples investigated Such discrepancies will be found in 
published data for other properties of copper 

6 A formula for the linear coefficient of expansion which applies in the range 16 C to 300 C is given 
below L is the length of the specimen at C, and L t its length at t C This formula was derived for 
hot-rolled copper rod containing 99.968% copper. 

LI = L (1 -f 00001623* + O.OOOOOQ00483t 2 , 

c The thermal conductivity decreases as the temperature rises, and the following values apply in the 
range C to 600 C. 



Thermal 
C Conductivity 

0.912 

20 0.910 

100 0.901 

200 890 



Thermal 

C Conductivity 

300 879 

400 867 

500 0856 

600 0.845 



4 Metals Handbook, 1936 ed., p. 1060, American Society for Metals, Cleveland. 
3 Handbook of Chemistry and Physics, 19th ed., op. cit., p. 1263. 



384 PROPERTIES OF COPPER 

Electrical Properties. The high electrical conductivity of copper, 
more than anything else, accounts for its widespread use. Only one 
metal, silver, has a higher electrical conductivity than copper; aluminum 
is the principal rival of copper for certain types of electrical conduc- 
tors, but in general copper is the best material available for carrying 
an electric current. 

Pure copper of electrolytic grade or its equivalent is the best for 
electrical use; many impurities greatly decrease the conductivity of 
copper, and there are no alloying elements which will improve it. 
Following is a list of the electrical resistivities of a few metals for 
comparison: 

Metal Resistivity, ohm/ cm? (18 to 20 C) 

Silver 1.629XKT 6 

Copper (annealed) 1.724 X 10~ 6 

Gold 2.44 X 10~ 6 

Aluminum 2.828 X 10~ 6 

Iron 10.0 X 10~ 6 

Lead 22 X 1(T 6 

The property which measures the resistance of a metal to the 
flow of electricity may be expressed as resistivity or as conductivity, 
which is the reciprocal of resistivity. The unit of resistivity is the 
ohm; the unit of conductivity is the reciprocal ohm (mho). The 
specific resistivity is the resistance in ohms of a section of standard 
dimensions; generally, as in the listing above, this is a piece 1 centi- 
meter long with a uniform cross-section of 1 square centimeter. But 
other dimensions may be used, as we shall see later. The resistance of 
a conductor increases directly with the length of the conductor, and 
inversely with its cross-section; thus the resistance of a copper bar 
1 meter long and 2 square centimeters in cross-section would be: 

1 724 X 1CP 6 

- - - - X 100 = 0.862 X 10- 4 = 0.0000862 ohm 
2i 

The conductivity of the same piece would be 



000862 
and the specific conductivity of copper would be 

1 10 6 

1.724 X 10^ = 1^4 =580,000 mho 

The conductivity of copper, however, is usually expressed in another 
fashion, namely as per cent of the resistance known as the International 



ELECTRICAL PROPERTIES 385 

Annealed Copper Standard. This value was adopted in 1913 to repre- 
sent the average conductivity of high-grade commercial conductivity 
copper, and copper having the same conductivity is said to have 100 per 
cent conductivity. If a given sample of copper has a conductivity of 
96.4 per cent, this means that its conductivity is 0.964 times that of 
the International Standard. Present day copper for conductivity 
purposes usually has a conductivity of 100 to 101 per cent, and 
occasional samples may run as high as 102 per cent. 

The example we have given before (1.724 X 10~ ohm per cm 3 ) is 
an example of volume resistivity. Resistivity and conductance may 
also be given in mass units; thus if we say that a sample of copper 
has a resistance of 0.15 ohm per meter-gram, we mean that 1 gram of 
the metal drawn into a uniform wire 1 meter long would have a 
resistance of 0.15 ohm. The density of commercial copper is subject 
to slight variations (as we have noted), and metal is bought by the 
pound rather than by the cubic foot; consequently the use of mass 
conductivity and resistivity is more common than the use of volume 
conductivity and resistivity. 

There are many different ways of expressing conductivity and 
resistivity; Table 2 gives a number of different values all these 
are equivalent to the International Annealed Copper Standard. In the 
parentheses the length of the conductor is given first; area or mass 
second. 

TABLE 2 4 * 6 

EQUIVALENTS OF THE INTERNATIONAL ANNEALED 
COPPER STANDARD AT 20 C 

Volume resistivity: 

1.7241 X 1(T 6 ohm (cm, cm 2 ) 
1.7241 microhm (cm, cm 2 ) 
0.67879 microhm (in , in. 2 ) 
10.371 ohm (ft, mil) 
0.017241 ohm (meter, mm 2 ) 

Volume conductivity: 

0.5800 megmho 6 (cm, cm 2 ) 
Mass resistivity: (density = 8.89) 

0.15328 ohm (meter, gram) 
875.20 ohm (mile, pound) 

1 microhm 10~* ohm. 
b 1 megmho = 10 6 mho. 

Motala Handbook, 1936 ed., p. 1063, American Society for Metals, Cleveland. 
Eahbach, O. W., op. cit., p. 11-94. 



386 PROPERTIES OF COPPER 

The electrical resistivity of copper varies with temperature, and 
from room temperature to the melting point of copper the variation is 
almost linear. An expression 4 for the mass resistivity of copper in 
the range to 150 C is 

R t = fl (l - 0.0041151* - 0.0000019988* 2 ) 

At extremely low temperatures copper (and other metals) exhibits 
super-conductivity , i.e., its electrical resistance almost vanishes. Some 
of the values for conductivity at low temperatures are: 3 

Temperature Resistivity, ohm (cm, cm 2 ) 

20 C 1.7241 X 1(T 6 

-100C 0.904 X KT 6 

- 206.6 C 0.163 X 1(T 6 

- 258.6 C 0.014 X 10~ 6 

Cold working increases the resistivity of copper, and the original 
conductivity is restored by annealing. 

CHEMICAL PROPERTIES 
Table 3 gives the chemical constants for the element copper. 

TABLE 3 3 ' 4 
CHEMICAL CONSTANTS FOR COPPER 

Atomic weight 63.57 

Atomic volume 7.11 

Atomic number 29 

Valence 1 or 2 

Electrochemical equivalent, Cu" 32940 mg/coulomb 

Cu' 65880 mg/coulomb 

Symbol, Cu (Latin Cuprum from Latin Cyprium, the Island of Cyprus) 
Position in periodic system Group I, Series, 5, Period 3 

Handbook of Chemistry and Physics, 19th ed , Chemical Rubber Publishing Co., Cleveland, 1934. 
Metals Handbook, 1936 ed , American Society for Metals, Cleveland. 

In the electromotive series copper lies below hydrogen and hence 
is not dissolved in dilute acids with the evolution of hydrogen. Iron, 
lead, tin, nickel, and zinc are all above copper in the series, and there- 
fore can be used to displace or cement copper from solutions of its 
salts. Platinum, palladium, gold, silver, and mercury are below copper 
(more " noble ") and can be displaced from solution by metallic copper. 

4 Metals Handbook, 1936 ed., American Society for Metals, Cleveland. 
3 Handbook of Chemistry and Physics, 19th ed., op. cit., p. 1338. 



GASES IN COPPER 



387 



The standard electrode potential for copper, as well as for a few of 
the common metals, is given in Table 4. 

TABLE 4 3 
ELECTRODE POTENTIALS 

[Metals Arranged m Order of Their Sequence in the Electromotive Force Series] 



Element 


Ion 


Electrode Reaction 


Electrode 
Potential 

(volts) 


Na 


Na+ 


Na ? Na + -f- e 


+2 7146 


Zn 


Zn++ 


|Zn ^ iZn++ + e 


+0 7618 


Fe 


Fe++ 


$Fe ^ |Fe++ + e 


40 441 


Ni 


NI++ 


|Ni ^Ni++ + e 


40 231 


H 2 


H+ 


fH 2 ^ H+ -f- e 


0.0000 


Cu 


Cu++ 


|Cu ^ |Cu++ + e 


-0.344 


Cu 


Cu+ 


Cu ^ Cu+ -f e 


-0 470 


Ag 


Ag+ 


Ag ^ Ag+ + e 


-0 7978 


Au 


Au f ~ H ~ 


^Au ^ | Au +++ -f e 


-1 360 



3 Handbook of Chemistry and Physics, 19th ed., p 850, Chemical Rubber Publishing Co , Cleveland, 
1934. 

Although copper does not displace hydrogen from acids, it readily 
dissolves in oxidizing acids (such as strong nitric acid) or in acids 
plus an oxidizing agent (e.g., H 2 S0 4 4 Fe 2 (S0 4 ) 8 ). Copper resists 
the action of the atmosphere and corrosive sea water and is widely 
employed for this reason. It does not usually give satisfactory service 
when exposed to acids, ammonia, or sulfur compounds. As we have 
noted, tough-pitch copper becomes embrittled when exposed to reducing 
gases at moderately high temperatures. 

Copper exposed to the atmosphere for a long period of time develops 
a bright green protective surface coating or patma of basic copper 
carbonate. Artificial patinas can be developed on copper surfaces 
by the action of special reagents. 

Gases in Copper. 6 Molten copper dissolves many of the common 
gases such as 2 , N 2 , C0 2 , CO, S0 2 , H 2 , and H 2 0. Many of these 
undoubtedly react with the copper or with other compounds, thus 
dissolved oxygen is probably in equilibrium with Cu 2 0: 

4Cu + 2 ;= 2Cu 2 

and reducing gases such as CO and H 2 react with any Cu 2 to form 
copper and C0 2 and H 2 0. 

6 Ellis, 0. W., A Review of Work on Gases in Copper: Am. Inst. Min. & Met. 
Eng. Trans., Vol. 106, p. 487, 1933. 



388 PROPERTIES OF COPPER 

The nature and amount of gas remaining in copper just before 
it solidifies has a marked effect on properties of the solid copper, as 
we have mentioned under " Fire Refining." 

There are, however, no reliable data on the solubility of various 
gases in copper which can be quickly summarized. This is a rather 
complex and difficult problem to investigate for several reasons. 
(1) It is necessary to work at high temperatures, which makes it dif- 
ficult to measure accurately small volumes of gas; (2) presence of even 
traces of impurities may have considerable effect on gas solubility; 
(3) many of the gases react with the copper so that it is not always a 
question of simple solubility. Commercial technique in copper con- 
verting and refining is based on more or less empirical rules, and a 
complete analysis of the effect of dissolved gases is still lacking. 

Solvents for Copper. Copper will dissolve in most acids when aided 
by oxidizing action, to form soluble copper salts. Copper also forms 
complex salts with ammonium compounds and with cyanides, and as 
we have seen, metallic copper is soluble in oxygenated ammonium car- 
bonate plus ammonia. The formula for cupric-ammonium carbonate 
we have given; other complex copper-ammonium salts are 

Cu(NH 3 ) 4 S04*H 2 0, cupric ammonium sulfate 
Cu(NH 3 ) 6 (OH) 2 , cuprammonium hydroxide 

In these salts the copper is in the form of a complex cuprammonium ion 
which has a deep-blue color. Cupric ions impart a greenish-blue color 
to the solution; cuprous ions are colorless. 

Examples of complex cyanides are the soluble sodium and potassium 
salts NaCu(CN) 2 and KCu(CN) 2 . In these salts the copper is 
present in the negative ion Cu(CN) 2 ~; these complex cyanides are 
used in electrolytes for electroplating copper. 

Cupric chloride forms a soluble double chloride with ammonium 
chloride CuCl 2 -2(NH 4 )Cl-2H 2 0. The insoluble cuprous chloride 
forms a similar compound with ferrous chloride, and this makes possible 
the dissolving of cuprous chloride in brines of ferrous chloride. 

EFFECTS OF MECHANICAL WORK ON PHYSICAL PROPERTIES 

In fabrication processes the wirebars, billets, or cakes are rolled down 
to rod or plates and these are further rolled into sheet (plates) or 
drawn into wire (rod). Other shapes are made by extruding metal 
through a die, by piercing billets to make seamless tubing, etc. All 
these processes are divided into two classes hot working and cold 
working. 



COLD WORKING 389 

Hot Working. Copper may be worked extensively at any tempera- 
ture up to about 1050 C. The higher the temperature the softer the 
metal and the less power required. The essential fact about hot work- 
ing is that the temperature of the metal must be kept above the re- 
crystallization temperature. When metal is deformed, the individual 
crystals are elongated and distorted, but if the temperature is high 
enough the atoms immediately reassemble to form new equi-axed 
grains. The size of the grains formed depends largely upon the tem- 
perature, and the higher the temperature the larger the crystal grains. 
The smallest grains will be formed just above the recrystallization 
temperature, and for this reason the finishing temperature of a hot- 
working operation should be around 400 or 500 C to avoid having an 
undesirable coarse grain in the finished object. Most hot-working 
operations begin with the metal at about 850 C. 

Hot working refines the coarse grains found in cast copper and 
increases the density of the metal by closing up small pores and gas 
holes. Otherwise it has little effect on the physical properties 
strength, ductility, electrical conductivity, hardness, thermal conduc- 
tivity, etc., remain about the same regardless of the amount of 
hot working. 

Cold Working. Cold working means the working of metal below the 
recrystallization temperature. Cold-worked copper is commonly 
known as " hard " copper, and cold working is the only method of 
hardening pure copper. The crystal grains in cold-worked copper 
are elongated, and the metal is harder and stronger than hot-worked 
or annealed copper. Cold working also decreases the electrical con- 
ductivity and ductility of copper. 

The effect of cold working on strength, hardness, and ductility is 
quite marked, as we have noted. The effect on electrical conductivity 
is less pronounced. The most severe cold working does not reduce the 
conductivity more than about 3 per cent. 

Cold working is usually a finishing operation ; in making cold-rolled 
sheet, for example, the rolling from the billet down almost to size 
would be done hot, and only the last few passes through the rolls 
would be made cold. The amount of hardening is determined by the 
per cent reduction of area of the piece during the cold rolling. Wire 
is made by cold drawing rod through a series of dies. 

Stamping, spinning, and drawing are all cold-working operations by 
means of which sheet copper is formed into a number of shapes. When 
working is severe (heavy " draft ") , oxygen-free copper is used be- 
cause of its superior ductility. 



390 



PROPERTIES OF COPPER 
HEAT TREATMENT OF COPPER 



Annealing is the only method of heat treatment used on pure copper, 
although other heat-treating processes may be used on some copper 
alloys. The purpose of annealing is to restore work-hardened copper 
to its original soft or ductile form. 



Silver, Per Cent 
0.034 0.068 0.103 0.137 





10 20 30 
Silver, Oz. per Ton 

(From Metals Handbook, p 1071, 1936 Ed.) 

FIG. 1. Effect of Silver on Anneal- 
ing Temperature of Cold-Rolled 
Copper Sheet. 



400 



800 



Soft 20 40 60 80 
Reduction by Drawing, % 

(From Metals Handbook, p. 1070, 1936 Ed ) 

FIG. 2. Characteristic Drawing 
Curves of Tough-Pitch Copper. 



1200 1600F. 



so 


1 






Ele 


ctrolytic 
LakeO 

Elonga 


.0005 Ag 
0868 Ag 

tion in 2 


1 i 


40 
20 


30 


r- 

n. 






*40 



~20 
( 




"""*, 


\ ' 


V 






V'j 




?"" 






* 


*., 


- 






i 




















i 
i 


7 




ensi 


leM 


engt 


h 








i 


' 




















) 200 400 600 800 10 



I 



Annealing Temp ,C, 
(From Metals Handbook, p. 1069, 1936 Ed.) 

FIG. 3. Effect of Silver on Annealing of Copper (0.05-inch sheet). 

The process of annealing involves the heating of the copper to the 
proper temperature, holding at this temperature for a certain period, 
and then allowing the metal to cool to room temperature. Most com- 
mercial annealing is done at 1100 F (590 C) which provides the 
necessary softening action without promoting undue grain growth. 
Theoretically it is only necessary to heat above the recrystallization 
temperature to obtain the full benefits of annealing, but practically 
it is necessary to use a somewhat higher temperature to make the 
process sufficiently rapid. Figure 3 shows the effect of various 
annealing temperatures on two specimens of work-hardened copper. 
Note that after a certain critical temperature is reached the tensile 



IMPURITIES IN COPPER 



391 



strength drops and the ductility rises. Annealing at higher tempera- 
tures has no further effect on the tensile strength, but the ductility 
drops somewhat because of the increased grain size. This diagram 
also shows the effect of silver on the recrystallization or annealing 
temperature of copper. 

For certain uses when slightly elevated temperatures are encountered 
in service such as engraver's plates, parts to be tinned or soldered, 
firebox plates and stays, etc., it is desirable that the part retain its 
work-hardened condition. Electrolytic copper would soften under 
these circumstances, so it is necessary to select a grade of copper with 
a high recrystallization temperature Lake copper or a synthetic 
silver-, arsenic-, or antimony-bearing copper. 

A light anneal is relatively short anneal designed principally to 
relieve mechanical strains in the cold-worked metal; it does not 
completely soften the metal. When the annealing is prolonged suf- 
ficiently to completely soften the metal, the operation is known as 
dead annealing. 




02 



06 .10 14 

Percentage of Impurity 



.18 



(Addicks in Metallurgy of Copper by Hofman and Hayward, McGraw-Hill Book Co., New York, 1924) 

FIG. 4. Effect of Impurities on the Conductivity of Copper. 

IMPURITIES IN COPPER 

Figure 4 shows the effect of some of the common impurities on the 
electrical conductivity of copper. Some elements show a pronounced 
lowering of the conductivity, whereas others have little effect. Oxygen 
in certain concentrations apparently increases tho conductivity, bub 
actually this is due to the fact that some of the harmful impurities are 
tied up as oxides and are not alloyed with the copper. Impurities affect 



392 PROPERTIES OF COPPER 

other properties of copper as well as the conductivity, and we shall 
briefly summarize the effects of the most common ones. 4 

Oxygen. Oxygen is present in all tough-pitch copper in the form 
of Cu 2 0. In cast copper it appears as the Cu-Cu 2 eutectic at the 
grain boundaries, and in worked copper it exists as small globules of 
Cu 2 visible under the microscope. In the concentrations found in 
commercial copper it has little effect on the mechanical properties. 

Sulfur, selenium, and tellurium are ordinarily regarded as harmful 
in copper, but they are actually dangerous only in tough-pitch copper 
unless their concentration exceeds 0.1 per cent. In the absence of 
oxygen these elements form the eutectics Cu-Cu 2 S, Cu-Cu 2 Se, and 
Cu-Cu 2 Te, which are similar in behavior and effect to the Cu-Cu 2 
eutectic. 

Silver in varying amounts is one of the most common impurities 
found in copper. It has a pronounced effect on the recrystallization 
temperature (Figs. 1 and 3), but otherwise it has little effect on the 
properties, and in specifications it is common to count silver as copper 
in the analyses. 

Bismuth is seldom encountered in detectable amounts in American 
coppers. When present in amounts over 0.001 per cent, however, it 
promotes brittleness of the metal and is very undesirable. Bismuth 
is not removed by fire refining alone but must be separated by 
electrolysis. 

Antimony is sometimes added to copper to raise the recrystallization 
temperature. In amounts up to 0.5 per cent it hardens the copper 
slightly and decreases the ductility, but it cannot be considered a 
harmful impurity unless the highest conductivity is desired. Antimony 
is extremely harmful to brass, so antimony-bearing scrap copper is 
not suited for brass making. 

Arsenic occurs naturally in Lake copper and is sometimes allowed to 
remain after refining in amounts up to 0.3 per cent or more. It has 
a small hardening and strengthening effect, and it raises the re- 
crystallization temperature of the metal. It decreases the electrical 
conductivity considerably. 

Iron is normally present in small amounts and is totally without 
effect on the mechanical properties of the metal ; it does, however, lower 
the electrical conductivity. 

Lead must not be present in amounts over 0.005 per cent if the 
copper is to be hot-rolled. Larger amounts, however, are without 
effect on the ductility at room temperature. The presence of oxygen 
decreases the harmful effect of lead. 

4 Metals Handbook, 1936 ed., p. 1072, American Society for Metals, Cleveland. 



COMMERCIAL CLASSES OF COPPER 393 

Cadmium is often added to copper in amounts of 0.7 to 1.0 per cent 
for the production of an alloy of high strength and good conductivity 
which is extensively used for trolley wire. In cold-drawn wire it is 
possible to obtain strengths as high as 92,000 pounds per square inch 
with a conductivity of 80 per cent International Annealed Copper 
Standard. Cadmium is rarely present in commercial copper. 

Phosphorus is almost never found in commercial copper, but a 
certain residual amount remains in deoxidized coppers when phosphorus 
is used as the deoxidizing agent. Phosphorus has such a detrimental 
effect on electrical conductivity that phosphorized copper is not suitable 
for electrical uses. 

COMMERCIAL CLASSES OF COPPER 

There are several types of commercial copper graded principally 
according to their chemical composition and electrical conductivity. 
Accepted universally by the trade are the specifications of the American 
Society for Testing Materials, and we shall present excerpts from 
several of the American Society for Testing Materials standards to 
illustrate the nature of these specifications. The important classes sold 
in America are (1) electrolytic, (2) Lake, and (3) fire-refined copper 
other than Lake, or casting copper. 

In addition to these there are other varieties such as Arsenical Lake, 
F H C, and Deoxidized, which are used for special purposes. In 
England there is a class known as Best Select which is a fire-refined 
copper that corresponds to the American Casting Copper. 

Electrolytic Cathode Copper. 7 Electrolytic cathode copper shall 
have a minimum purity of 99.90 per cent, silver being counted as 
copper. The copper shall have a resistivity not to exceed 0.15436 in- 
ternational ohms per meter-gram at 20 C (annealed). Cathodes 
shall be hard enough to stand ordinary handling without excessive 
breakage or excessive separation of nodules, and shall be substantially 
free from all foreign material, for example, copper sulfate, dirt, grease, 
and oil. 

Electrolytic Copper Wirebars, Cakes, Slabs, Billets, Ingots, and Ingot 
Bars. 8 The copper in all shapes shall have a purity of at least 99.900 
per cent, silver being counted as copper. 

All wirebars shall have a resistivity not to exceed 0.15436 interna- 
tional ohms per meter-gram at 20 C (annealed) ; all ingots and ingot 
bars shall have a resistivity not to exceed 0.15694 international ohms 
per meter-gram at 20 C (annealed). 

7 A.S.T.M. Designation, B 116-38 T. 

8 A.S.T.M. Designation, B 5-27. 



394 PROPERTIES OF COPPER 

Cakes, slabs, and billets shall come under the ingot classification, 
except when specified for electrical use at time of purchase, in which 
case wirebar classification shall apply. 

Wirebars, cakes, slabs, and billets shall be substantially free from 
shrink holes, cold sets, sloppy edges, concave tops, and similar defects 
in set or casting. This clause shall not apply to ingots or ingot bars, 
in which case physical defects are of no consequence. 

Lake Copper Wirebars, Cakes, Slabs, Billets, Ingots, and Ingot 
Bars. 9 In order to be classed as Lake, copper must originate on the 
northern peninsula of Michigan, U. S. A. 

Lake copper offered for electrical purposes, whether fire or electro- 
lytically refined, shall be known as Low Resistance Lake; Lake copper 
having a resistivity greater than 0.15694 international ohms per meter- 
gram at 20 C shall be known as High Resistance Lake. 

Low Resistance Lake. Wirebars shall have a resistivity not to 
exceed 0.15436 international ohms per meter-gram at 20 C (annealed). 
Ingots and ingot bars shall have a resistivity not to exceed 0.15694 
international ohms per meter-gram at 20 C (annealed) . Cakes, slabs, 
and billets shall come under the ingot classification except when 
specified for electrical use at time of purchase, in which case wirebar 
classification shall apply. 

The purity shall be at least 99.900 per cent as determined by electro- 
lytic assay, silver being counted as copper. 

High Resistance Lake. The purity shall be at least 99.900 per cent, 
copper, silver, and arsenic being counted together. The arsenic content 
of High Resistance Lake copper, when required for special purposes, 
shall be the subject of agreement at time of purchase. 

Fire-Refined Copper Other than Lake. 10 These specifications cover 
fire-refined copper other than Lake and not usually electrolytically 
refined. Fire-refined copper other than Lake is intended for use in 
rolling into sheets and shapes for mechanical purposes and is not in- 
tended for electrical purposes nor wrought alloys. 

9 A S T M. Designation, B 4-27. 
10 A.S.T.M. Designation, B 72-33. 



COPPER BARS FOR LOCOMOTIVE STAYBOLTS 395 

The copper in all shapes shall conform to the following chemical 
composition: 

Per Cent 

Copper plus silver, minimum 99.7000 

Arsenic, maximum 0.1000 

Antimony, do 0.0120 

Bismuth, do 0.0020 

Iron, do 0100 

Lead, do 0.0100 

Nickel, do 0.1000 

Oxygen, do 0750 

Selenium, do 0.0400 

Tellurium, do 0.0140 

Tin, do 0.0500 

Copper Bars for Locomotive Staybolts. 11 Copper bars for loco- 
motive staybolts will serve as an example of the specifications required 
for various types of fabricated and semi-fabricated copper objects. 

These specifications cover two grades of copper bars for locomotive 
staybolts, namely, Arsenical and N on- Arsenical. The copper shall be 
fire-refined or electrolytic and shall be finished to dimensions by hot- 
rolling from suitable bars. 

The copper shall conform to the following chemical composition: 

Arsenical Copper shall contain 0.25 to 0.50 per cent arsenic and shall 
contain not more than 0.120 per cent of impurities exclusive of arsenic 
and silver. 

Non-Arsenical Copper shall have a purity of at least 99.90 per cent, 
silver being counted as copper. The total of impurities other than 
silver shall not exceed 0.10 per cent. 

The material shall conform to the following minimum requirements 
as to tensile properties: 

Arsenical Non-Arsenical 

Tensile strength, psi 31,000 30,000 

Elongation in 8 inches, per cent 35 30 

The test specimen shall stand being bent cold through 180 flat on 
itself without cracking on the outside of the bent portion. One tension 
and one bend test shall be made from each 5000 pounds or fraction 
thereof. 

The bars shall be truly round within 0.01 inch and shall not vary 
more than 0.01 inch over nor more than 0.005 inch under the specified 
size. 

n A.S.T.M. Designation, B 12-33. 



CHAPTER XI 

THE USES OF COPPER 

INTRODUCTION 

In this chapter we shall give a brief discussion of the various uses 
of metallic copper and copper alloys. 

Table 1 gives the estimated use of copper in the United States for 
the period 1933-1937. This includes all copper produced, both primary 
and secondary (scrap), and includes all copper used, both as pure 
metal and as alloys. For 1937 the percentage for each classification is 
computed from the data in the preceding column. It is almost im- 
possible to set up a table giving the " average " consumption of copper 
over a period of years, as consumption varies considerably from year 
to year as to both totals and distribution. Table 1 shows several 
examples of this. 

1. The total consumption of copper in 1937 was more than twice 
the 1933 consumption. 

2. Consumption of copper for electrical manufactures, telephones 
and telegraph, and light and power lines was also about twice as 
great in 1937 as in 1933. 

3. Air conditioning showed a decided increase from 1934 to 1937, 
with no indicated consumption in 1933. 

These and many other facts show that the consumption of copper 
depends upon general economic conditions and technological trends, 
but it will not be profitable to devote more space at this point to a 
further examination of consumption statistics. 

Approximately 60 per cent of the copper consumed is used as 
metal; the remaining 40 per cent is used in alloys. About 38 per cent 
(Table 1) of the entire copper consumption is used for various 
electrical purposes, and this represents about 63 per cent of the total 
metallic copper used (excluding alloys) . 

Reasons for Using Copper. Copper is not used as pure metal 
because of its mechanical properties; pure copper is a relatively soft 
and weak metal and has a high specific gravity. Where strength is 
the principal requirement there are other metals and alloys which are 
much more suitable. This restriction of course, does not apply to 
copper alloys, many of which exhibit exceptional mechanical properties. 

396 



REASONS FOR USING COPPER 



397 



TABLE 1 

ESTIMATED USE OF COPPER IN THE UNITED STATES, 
1933-1937, IN SHORT TONS 





1933 


1934 


1935 


1936 


1937 


1937 


Electrical manufactures^ 


90,000 


101,000 


128,000 


164,000 


213,000 


24 8 


Telephones and tele- 














graphs 


18,000 


18,000 


18,000 


26,000 


30,000 


3 5 


Light and power lines 


33,000 


36,000 


55,500 


72,000 


83,000 


9 6 


Wire cloth 


5,000 


4,600 


5,600 


6,500 


6,600 


8 


Other rod and wire 


46,000 


40,000 


48,000 


90,000 


112,000 


13 


Ammunition 


10,500 


13,500 


13,700 


11,900 


14,100 


1 6 


Automobiles'* 


49,000 


63,000 


95,000 


108,000 


112,000 


13 


Buildings 6 


36,000 


36,000 


49,000 


71,000 


70,500 


8 2 


Castings, n.e.s ' 


36,000 


36,000 


36,000 


39,000 


40,000 


4 6 


Clocks and watches 


2,800 


2,200 


2,400 


3,400 


3,200 


4 


Coinage 


100 


900 


1,500 


2,000 


100 





Copper-bearing steel 


1,500 


2,100 


2,300 


3,900 


4,600 


5 


Fire-fighting apparatus 


1,100 


1,000 


1,200 


1,300 


1,500 


2 


Radiators, heating 


2,400 


1,000 


1,100 


2,000 


2,100 


2 


Radio receiving sets 


11,500 


12,500 


16,000 


24,000 


23,100 


2 7 


Railway equipment^ 


800 


2,100 


1,800 


4,000 


7,100 


8 


Refrigerators* 


11,400 


15,700 


15,400 


15,000 


13,500 


1.6 


Shipbuilding* 


1,800 


3,200 


1,100 


5,000 


6,400 


7 


Washing machines* 


1,000 


1,400 


1,300 


1,500 


1,500 


2 


Water heaters, household 


1,500 


1,500 


1,500 


1,500 


1,500 


2 


Air conditioning*** 





3,800 


4,800 


6,400 


7,200 


0.8 


Other uses 


40,000 


42,000 


46,000 


59,000 


62,000 


7.2 


Manufactures for export 


15,600 


25,500 


29,500 


31,600 


45,000 


5 2 




415,000 


463,000 


574,000 


749,000 


860,000 


99 8 



a Minerals Yearbook, 1938, p 93, U. S Bur. Mines. 

b Generators, motors, electric locomotives, switchboards, light bulbs, etc. 

c Transmission and distribution wire and busbars, accounting only for the public utility companies 

d Does not include starter, generator, and ignition equipment 

* Excludes electrical work. 

f Bearings, bushings, lubricators, valves, and fittings. 

Includes air conditioning 

h Exclusive of electrical equipment. 

* Other than railway. 

The principal reasons for the widespread industrial use of metallic 
copper are (1) electrical conductivity, (2) heat conductivity, (3) ductil- 
ity, (4) resistance to corrosion, and (5) decorative value. 

The first of these is by far the most important, and more than 
anything else accounts for the importance of copper. The second is 
likewise very important and accounts for the use of copper in re- 
frigerators, radiators, water heaters, air conditioning, etc., where the 



398 THE USES OF COPPER 

rapid transfer of heat is essential. These two properties are peculiarly 
important because pure copper is exceeded in electrical and thermal 
conductivity only by the relatively expensive metal silver. Moreover, 
all alloying elements added to copper appear to lower both the electrical 
and thermal conductivity, and there seems to be small likelihood that 
any alloy or other material will be discovered which will be superior to 
copper in these two respects. 

Copper is a very ductile metal and for that reason is suited for 
purposes which require extensive cold working wire drawing, cold 
stamping, spinning, deep drawing, etc. Oxygen-free copper is even 
more ductile than tough-pitch copper and will stand more extensive 
cold working than any other metal or alloy with the exception of some 
of the precious metals. 

The resistance of copper to corrosion by certain reagents is responsible 
for some of its uses, particularly in the manufacture of vessels for 
holding corrosive liquids, tubes for conveying corrosive liquids, and 
sheathing for boats. This use, however, is conditioned by two other 
factors. There are other metals and alloys which are superior to pure 
copper as corrosion resistant materials, and when copper is used it is 
because (1) it is cheaper than the other material, or (2) its superior 
ductility makes it easier to form the required shape. The latter con- 
sideration applies particularly to the manufacture and use of small- 
bore tubing. Of course there are some corrosive materials which 
readily attack copper and preclude its use in contact with them. 

Finally there is the fact that the color and luster of copper make it 
desirable for its decorative effect. The green patina formed on copper 
exposed to the atmosphere gives an attractive color to copper roofing. 

The reasons listed above, singly or in combination, account for 
practically all of the industrial uses made of commercially pure 
copper; the use of copper in alloys is dictated by many other con- 
siderations. 

Temper. We shall have occasion in the discussion of copper and 
copper alloys to refer to the temper of a particular sample. As applied 
to non-ferrous metals and alloys, temper means the condition of the 
metal or alloy with respect to its previous mechanical and/or heat 
treatment. There is no standard method for designating temper, but 
the following examples will indicate some common usages: 

1. Soft copper may be designated as soft annealed, dead soft, dead 
annealed, etc. 

2. Cold-worked copper is known as hard, cold drawn, cold rolled, 
hard drawn, full hard f etc. 



ELECTRICAL CONDUCTORS 399 

These terms generally refer to completely hardened copper; if the 
reduction in area has been only enough to partly harden the copper, 
then the metal is medium hard, half hard, medium hard-drawn, etc. 
In some coppers the amount of cold working is indicated ; a designation 
which might be used for copper sheet is "0.0325 gauge, 3 numbers 
hard," which means that after the last anneal the metal was reduced by 
cold rolling 3 Brown and Sharpe gauge numbers to a thickness of 0.0325 
inch. 

A similar terminology applies to alloys, except that some alloys can 
be hardened by heat treatment, and the terms used must indicate 
whether the hard state was produced by cold working or by heat 
treatment. 



ELECTRICAL CONDUCTORS 

We have already considered the question of the electrical conductivity 
of copper, and at this point we shall briefly discuss the commercial 
shapes of copper used as conductors. 

Busbars, Etc. For carrying heavy currents over short distances 
copper busbars or busses are used. These must have a cross-section 
large enough to carry the current required. Busses are made in dif- 
ferent cross-sectional shapes; they may be rectangular, triangular, or 
cylindrical bars, I-beams, angles, or pipe. Generally busses are used 
bare, and the nature of the insulating supports will depend on the 
voltage drop between the bus and its surroundings. 

Wire. Conductivity wire is made in a variety of sizes and con- 
ditions. Wires are usually circular in cross-section. They may be 
bare or covered with one of a number of insulating substances; a 
" wire " may be a single wire or it may be " stranded " or made up of 
a number of smaller wires. Large " cables " may consist of hundreds 
of small wires each insulated from the others, and some of these are 
sheathed in metallic load over the insulation. A number of insulating 
materials may be used depending upon the use for which the wire is 
intended. These include special varnishes and enamels, fabrics, rubber, 
and asphalt. 

In designating the sizes of individual wires, several special units are 
employed in addition to the common English and metric units. 

Special Units. A common unit for the diameter of wire is the mil, 
which is one one-thousandth of an inch. A square ,nil is the area of a 
square 1 mil on a side, and a circular mil is the area of a circle 1 mil 



400 THE USES OF COPPER 

in diameter. These may be converted to other units by the following 
relations: 

1 inch = 1000 mils 

1 millimeter = 39.37 mils 

1 centimeter =- 393.7 mils 

1 square millimeter = 1973 circular mils 

1 square centimeter = 1.973 X 10 5 circular mils 

1 square inch = 1.273 X 10 6 circular mils 

1 square mil == 1.2732 circular mils 

Wires larger than l / 2 i nc h are usually designated by the wire diameter 
in mils or other units, but smaller wires have their size denoted by 
certain arbitrary " gauge numbers." The larger the gauge number 
the smaller the diameter of the wire. The principal wire gauge used 
in the United States is the American, or Brown and Sharpe gauge, 
which is abbreviated A.W.G. or B. and S. 

The ratio of the diameter of any wire to the next smaller wire in the 
B. and S system is \/92 to 1 or approximately 1.1229 to 1. Thus 
No. 36 B. and S. wire has a diameter of 5 mils, and the diameter of 
the next larger wire (No. 35 B. and S.) is 5.0 X 1.1229 = 5.61 mils. 
The surd ^/92 is approximately equal to \/2 ( = 1.1225), and this 
makes it possible to have a group of wires of regular gauge size with an 
aggregate area approximately equal to that of another regular gauge 
size. Thus a reduction of three gauge numbers (say from No. 36 to 
No. 33 B. and S.) results in a new gauge number representing a 
diameter approximately \/2 times that of the original gauge number 
or an area about twice as great. In other words the No. 33 wire has 
twice the cross-sectional area of the No. 36 wire. 

From the definitions given above, the following approximate rela- 
tions may be derived : 

An increase of 1, 2, and 3 in the number increases the resistance 25, 
60, and 100 per cent, respectively. An increase of 10 in the number 
increases the resistance 10 times. The cross-sectional area and 
weight per foot will vary inversely with the resistance, and by taking 
the constants for one wire, the approximate values for the other wires 
can be calculated. No. 10 B. and S. wire is convenient for a starting 
point. Its approximate characteristics are: 

Ohms per 1000 feet - 1 
Circular mils area 10,000 
Weight, pounds per 1000 feet - 32 



WIRE 401 

Thus the resistance of No. 12 B. and S. wire would be 1 X (1.0 + 0.60) 

10 000 

= 1.6 ohms; the area would be - = 6250 circular mils; the diame- 

1.6 

00 

terV6250 = 79 mils; and the weight = 20 pounds per 1000 feet. 

B. and S. numbers range from No. 0000 (diameter 460 mils) to 
No. 40 (diameter = 3.1 mils). 

Copper wire may be hard-drawn, medium hard-drawn, or annealed, 
and standard specifications for various size wires are shown in Tables 
2, 3, 4, and 5. 



402 



THE USES OF COPPER 



TABLE 2 fl 

WIRE TABLE, STANDARD ANNEALED COPPER AMERICAN WIRE GAGE (B. & S.). 

ENGLISH UNITS 



&0 

P 


Diam- 
eter 
m Mils 
at20C 


Cross-section at 20 C 


Ohms per 
1000ft 6 
at 20 C 

(= 68 F) 


Pounds 
per 
1000ft 


Feet 
per 
Pound 


Feet per 
Ohm c at 
20 C 
(=68F) 


Ohms per 
Pound at 
20 C 
(= 68 F) 


Pounds 
per 
Ohm at 20 C 

(= 68 F) 


Circular 
Mils 


Square 
Inches 


0000 


4600 


211600 


01662 


049 01 


6405 


1 561 


20400 


000 076 52 


13 070. 


000 


4096 


167800. 


1318 


.06180 


5079 


1968 


16180 


0001217 


8219. 


00 


3648 


133100 


1045 


.077 93 


4028 


2482 


12830 


0001935 


5169. 





3249 


105500 


08289 


09827 


3195 


3130 


10180 


0003076 


3251. 


1 


2893 


83690 


06573 


1239 


2533 


3947 


8070 


1 0004891 


2044. 


2 


2576 


66370 


05213 


1563 


2009 


4977 


6400 


, 0007778 


1286 


3 


2294 


52640 


04134 


1970 


1593 


6276 


5075 


' 001237 


8086 


4 


2043 


41 740 


032 78 


2485 


1264 


7914 


4025 


' 001966 


508.5 


5 


1819 


33100 


02600 


3133 


1002 


9980 


3192 


003127 


3198 


6 


1620 


26250 


02062 


3951 


7946 


1258 


2531 


004972 


2011 


7 


1443 


20820 


01635 


4982 


6302 


1587 


2007 


. 007905 


1265 


8 


1285 


16510 


01297 


6282 


4998 


2001 


1592 


01257 


7955 


9 


1144 


13090 


01028 


7921 


3963 


2523 


1262 


i 01999 


5003 


10 


1019 


10380 


008155 


9989 


31 43 


3182 


1001 


03178 


3147 


11 


9074 


8234 


006467 


1260 


2492 


4012 


7940 


05053 


1979 


12 


8081 


6530. 


005 129 


1588 


1977 


5059 


6296 


08035 


1245 


13 


7196 


5178 


004067 


2003 


1568 


6380 


4993 


1278 


7827 


14 


6408 


4107 


003225 


2525 


1243 


8044 


39bO 


2032 


4922 


15 


5707 


3257 


002558 


3184 


9858 


1014 


3140 


3230 


3096 


16 


5082 


2583 


002028 


4016 


7818 


1279 


2490 


5136 


1947 


17 


4526 


2048 


001609 


5064 


6200 


1613 


1975 


8167 


1224 


18 


4030 


1624 


001276 


6385 


4917 


2034 


1566 


1299 


07700 


19 


3589 


1288 


001012 


8051 


.3899 


2565 


1242 


2065 


4843 


20 


3196 


1022 


000 802 3 


1015 


3092 


3234 


9850 


3283 


3046 


21 


2846 


8101 


0006363 


1280 


2452 


4078 


7811 


|5221 


1915 


22 


2535 


6424 


0005046 


1614 


1945 


5142 


6195 


8301 


1205 


23 


2257 


5095 


0004002 


2036 


1542 


6484 


4913 


1320 


07576 


24 


2010 


4040 


0003173 


2567 


1223 


8177 


3896 


2099 


04765 


25 


1790 


3204 


0002517 


3237 


09699 


1031 


3090 


3337 


02997 


26 


1594 


2541 


0001996 


4081 


.7692 


1300 


2450 


5306 


01885 


27 


1420 


2015 


000 158 3 


5147 


6100 


1639 


1943 


8437 


01185 


28 


1264 


1598 


0001255 


6490 


4837 


2067 


1541 


1342 


007 454 


29 


1126 


1267 


00009953 


8183 


3836 


2607 


1222 


2133 


004688 


30 


1003 


1005 


000 078 94 


1032 


.3042 


3287 


9691 


3392 


002948 


31 


8928 


7970 


000 062 60 


1301 


.2413 


4145 


7685 


5393 


.001 854 


32 


7950 


6321 


000 049 64 


164 1 


1913 


5227 


6095 


8576 


001 166 


33 


7080 


5013 


000 039 37 


2069 


1517 


6591 


4833 


1364 


0007333 


34 


6305 


3975 


00003122 


2609 


1203 


8310 


3833 


2168 


0004612 


35 


5615 


3152 


000 024 76 


3290 


09542 


10480 


3040 


3448 


0002901 


36 


5000 


2500 


000 019 64 


4148 


07568 


13 210 


2411 


5482 


0001824 


37 


4453 


1983 


00001557 


5231 


06001 


16660 


1912 


8717 


0001147 


38 


3965 


1572 


00001235 


6596 


04750 


21010 


1516 


13860 


000 072 15 


39 


3531 


1247 


000009793 


8318 


03774 


26500 


1202 


22040 


00004538 


40 


3145 


9888 


000007766 


1049 


02993 


33410 


09534 


35040 


00002854 



Eshbach, W O , Handbook of Engineering Fundamentals, p 11-93, John Wiley <fe Sons, 
New York, 1936 

b Resistance at the stated temperatures of a wire whose length is 1000 ft at 20 deg cent. 
c Length at 20 deg cent of a wire whose resistance is 1 ohm at the stated temperatures 



WIRE 



403 



TABLE 3 a 
SPECIFICATIONS FOR HARD-DRAWN AND MEDIUM HARD-DRAWN COPPER WIRE 







Hard-drawn Copper Wire b 


Medium Hard-drawn Copper Wire c 


Diameter, 
in 


Area, 
cir mils 


Tensile 
Strength, 


Elongation, 
per cent 


Tensile Strength, 
Ib per sq in 


Elongation, 
per cent 






Ib per sq in 


in 60 in 


Minimum 


Maximum 


in 60 in 


460 


211,600 


49,000 


3 75 d 


42,000 


49,000 


3 75 d 


410 


168,100 


51,000 


3 25 d 


43,000 


50,000 


3 6 d 


3G5 


133,225 


52,800 


2 80 d 


44,000 


51,000 


3 25 d 


325 


105,625 


54,500 


2 40 d 


45,000 


52,000 


3 O d 


289 


83,520 


56,100 


2 I7 d 


46,000 


53,000 


2 75* 


258 


66,565 


57,600 


98 d 


47,000 


54,000 


2 5 d 


229 


52,440 


50,000 


79 rf 


48,000 


55,000 


2 25 d 


204 


41,615 


60,100 


24 


48,330 


55,330 


1 25 


182 


33,125 


61,200 


.18 


48,600 


55,660 


1 20 


165 


27,225 


62,000 


14 








162 


26,245 


62,100 


14 


49,000 


56,000 


1 15 


144 


20,735 


63,000 


09 


49,330 


56,330 


1 11 


134 


17,956 


63,400 


07 








128 


16,385 


63,700 


06 


49,660 


56,660 


1 08 


114 


12,995 


64,300 


02 


50,000 


57,000 


1 06 


104 


10,815 


64,800 


1 00 








102 


10,404 


64,900 


1 00 


50,330 


57,330 


1 04 


092 


8,464 


65,400 


97 








091 


8,281 


65,400 


97 


50,660 


57,660 


1 02 


081 


6,561 


65,700 


95 


51,000 


58,000 


1 00 


080 


6,400 


65,700 


94 








072 


5,184 


65,900 


92 


51,330 


58,330 


98 


065 


4,225 


66,200 


91 








064 


4,096 


66,200 


90 


51,660 


58,660 


96 


057 


3,249 


66,400 


89 


52,000 


59,000 


94 


051 


2,601 


66,600 


87 


52,330 


59,330 


92 


045 


2,025 


66,800 


86 


52,660 


59,660 


90 


040 


1,600 


67,000 


85 


53,000 


60,000 


88 






Maximum resistivity at 


Maximum resistivity at 20 C: for 






20 C for diameters 


diameters 460 to 325 in , 896 15 Ib per 






460 to 325 in , 900 77 


mile-ohm (10 b!9 ohms (mil, ft)); for 






Ib per mile-ohm (10674 


diameters 324 to 040 in , 905 44 Ib 






ohms (mil, ft)) , for diam- 


per mile-ohm (10 729 ohms (mil, ft)). 






eters 324 to 040 in , 








910 15 Ib per mile-ohm 








(10 785 ohms (mil, ft)) 





a Eshbach, Handbook of Engineering Fundamentals, p 11-94, John Wiley & Sons, New York, 1936. 
b A.S.T.M Standard Bl-27, A.S.A. Standard H 14-1929. 
A S T.M. Standard B2-27 
d Elongation per cent in 10 in 



404 



THE USES OF COPPER 



TABLE 4 

SPECIFICATIONS FOR SOFT OR ANNEALED COPPER WIRE AND TINNED SOFT OR 
ANNEALED COPPER WIRE FOR RUBBER INSULATION 





Soft or Annealed Copper Wire 6 


Tinned Soft or Annealed Copper Wire 
for Rubber Insulation 


Diameter 

(in ) 


Tensile 


Elonga- 


Maximum 


Tensile 


Elonga- 


Maximum 




Strength 


tion in 


Resistivity d 


Strength 


tion in 


Resistivity d 




(Ib/sq in ) 


10 in 


(Ib/nule-ohm) 


max (Ib/sq 


10 in. nun 


(Ib/mile-ohm) 






(per cent) 




in) 


(per cent) 




460 to 0.290 


36,000 


35 


891 58 


36,000 


30 


896.15 


0.289 toO 103 


37,000 


30 


891 58 


37,000 


25 


900.77 


102 to 021 


38,500 


25 


891 58 


38,500 


20 


910 15 


0.020 toO 012 


40,000 


20 


891 58 


39,000 


15 


929.52 


Oil toO 003 


40,000 


20 


891 58 


40,000 


10 


939.51 



Eshbach, O. W., Handbook of Engineering Fundamentals, p 11-95, John Wiley and Sons, Inc., 
New York, 1936 

6 A.S.T.M. Standard B3-27, A.S.A. Standard H4-1928 and C862-1928; A.I.E E. Standards 60, 
61-1928. 

C AS.TM Standard B33-21; A.S.A. Standard H16-1928 and C861-1928, A.I.E.E. Standards 
60, 61-1928 

d At 20 C (68 F). 



WIRE 



405 



TABLE 5 a 

ALLOWABLE CARRYING CAPACITIES OF COPPER WIRES* 
(NATIONAL ELECTRICAL CODE) 



Gauge 
No. 
A.W.G. 


Diameter of 
Solid Wires 
(mils) 


Area 
(cir mils) 


Rubber 
Insulation 
(amp) 


Varnished 
Cambric 
Insulation 
(amp) 


Other 
Insulation 
(amp) 


18 


40.3 


1,624 


3 


. . 


5 C 


16 


50.8 


2,583 


6 


. . 


10 


14 


64.1 


4,107 


15 


18 


20 


12 


80.8 


6,530 


20 


25 


30 


10 


101.9 


10,380 


25 


30 


35 


8 


128 5 


16,510 


35 


40 


50 


6 


162.0 


26,250 


50 


60 


70 


5 


181.9 


33,100 


55 


65 


80 


4 


204.3 


41,740 


70 


85 


90 


3 


229.4 


52,630 


80 


95 


100 


2 


257.6 


66,370 


90 


110 


125 


1 


289.3 


83,690 


100 


120 


150 





325.0 


105,500 


125 


150 


200 


00 


364.8 


133,100 


150 


180 


225 


000 


409.6 


167,800 


175 


210 


275 




.... 


200,000 


200 


240 


300 


0000 


460.0 


211,600 


225 


270 


325 






250,000 


250 


300 


350 






300,000 


275 


330 


400 






350,000 


300 


360 


450 






400,000 


325 


390 


500 






500,000 


400 


480 


600 






600,000 


450 


540 


680 






700,000 


500 


600 


760 






750,000 


525 


630 


800 






800,000 


550 


660 


840 







900,000 


600 


720 


920 







1,000,000 


650 


780 


1,000 






1,100,000 


690 


830 


1,080 






1,200,000 


730 


880 


1,150 






1,300,000 


770 


920 


1,220 






1,400,000 


810 


970 


1,290 






1,500,000 


850 


1,020 


1,360 






1,600,000 


890 


1,070 


1,430 






1,700,000 


930 


1,120 


1,490 






1,800,000 


970 


1,160 


1,500 




. . 


1,900,000 


1,010 


1,210 


1,610 


.... 




2,000,000 


1,050 


1,260 


1,670 



Eahbach, O. W., Handbook of Engineering Fundamentals, p. 11-95, John Wiley and Sons, 
Inc., New York, 1936. 

6 Copper wires and cables of 98 per cent conductivity. For aluminum wire the allowable carrying 
capacities shall be taken as 84 per cent of those given in the table for the respective sizes of copper wire 
with the same kind of covering. 

e The allowable carrying capacities of No. 18 and No. 16 are 10 and 15 amp, respectively, when in 
oords for portable heaters, types HC and HPD. 



406 THE USES OF COPPER 

Aluminum Conductors. Aluminum conductors compete with copper 
principally in large conductors where the saving in weight is of im- 
portance. The largest use for aluminum conductors is for power 
transmission through cables having a reinforcing core of high-strength 
steel wire. These composite cables are larger than copper cables of 
the same conductivity but are lighter in weight; at very high voltages, 
such as 100,000 volts, the greater diameter results in a lower corona 
loss. All-aluminum cables are used for railway feeders to carry 
heavy currents at low voltages, and aluminum busbars are used for 
both switchboards and general power transmission. 

Hard-drawn aluminum wire has a conductivity of about 61 per 
cent, I.A.C.S. Thus the volume conductivity of aluminum is only 
about 61 per cent of that of copper. The mass conductivity, how- 
ever, is about twice that of copper, as the specific gravity of aluminum 
is 2.7 and copper has a density of 8.89. For copper and aluminum 
wires of equal resistance per unit length, the following ratios apply: 

Copper Aluminum 
Cross-section 1 1.61 

Diameter 1 1.27 

Weight 1 0.488 

Breaking strength 1 0.64 

Thus we see that for equal conductivity per unit length, the aluminum 
conductor has only about one-half the weight of the copper; however, 
the aluminum wire is larger and has less strength. 

Aluminum wire has been used for some motor windings, but these 
windings must occupy more space than copper windings if the same 
operating temperature is to be maintained (same total resistance). 

OTHER WIRE AND ROD 

Not all copper rod and wire are used in electrical work, but are used 
for other purposes, such as copper bars for locomotive staybolts (rod) 
and woven-wire screen or cloth. 

COPPER SHEET AND STRIP 

Copper sheet and strip is rolled from cakes of tough-pitch or oxygen- 
free copper. Copper sheets may be obtained in integral multiples of 
1/16 of an inch up to 2 inches in thickness. The sheets thinner than 
about y 2 inch may be designated by gauge numbers; the B. and S. 
wire gauge is most commonly used for this purpose, and the numbers 
have the same significance as in the case of wire; i.e., No. 0000 sheet 
would be 0.4600 inch thick. Table 6 gives the gauge numbers, thick- 



COPPER SHEET AND STRIP 



407 



TABLE 6 a 

AMERICAN WIRE GAUGE AND WEIGHTS OP COPPER, ALUMINUM, AND 
* BRASS SHEETS AND PLATES 



Gauge 
No. 


Thickness 


Approximate weight 6 Ib/sq ft 


Inch 


mm 


Copper 


Aluminum 


Commercial 
(high) brass 


0000* 


4600 


11 68 


21 27 


6 49 


20 27 


000 


4096 


10 40 


18 94 


5.78 


18 05 


00 


3648 


9 266 


16 87 


5 14 


16 07 





3249 


8 252 


15 03 


4 58 


14 32 


1 


2893 


7 348 


13 38 


4 08 


12 75 


2 


2576 


6 544 


11 91 


3 632 


11 35 


3 


2294 


5 827 


10 61 


3 234 


10 11 


4 


2043 


5 189 


9 45 


2 880 


9 00 


5 


1819 


4 621 


8 41 


2 565 


8 01 


6 


1620 


4 115 


7 49 


2 284 


7 14 


7 


1443 


3 665 


6 67 


2 034 


6 36 


8 


1285 


3 264 


5 94 


1 812 


5 66 


9 


1144 


2 906 


5 29 


1 613 


5 04 


10 


1019 


2 588 


4 713 


1 437 


4 490 


11 


0907 


2 305 


4 195 


1 279 


3 996 


12 


0808 


2 053 


3 737 


1 139 


3 560 


13 


0720 


1 828 


3 330 


1 015 


3 172 


14 


0641 


1 628 


2 965 


904 


2 824 


15 


0571 


1 450 


2 641 


805 


2 516 


16 


0508 


1 291 


2 349 


716 


2 238 


17 


0453 


1 150 


2 095 


639 


1 996 


18 


0403 


1 024 


1 864 


568 


1 776 


19 


0359 


9116 


1 660 


506 


1 582 


20 


0320 


8118 


1 480 


451 


1.410 


21 


0285 


7230 


1 318 


402 


1 256 


22 


0253 


6438 


1.170 


3567 


1 115 


23 


0226 


5733 


1 045 


3186 


996 


24 


0201 


5106 


930 


2834 


886 


25 


0179 


4547 


828 


2524 


789 


26 


0159 


4049 


735 


2242 


701 


27 


0142 


3606 


657 


2002 


626 


28 


0126 


3211 


583 


1776 


555 


29 


0113 


2859 


523 


1593 


498 


30 


0100 


2546 


4625 


1410 


4406 


31 


00893 


2268 


4130 


1259 


3935 


32 


00795 


2019 


3677 


1121 


3503 


33 


00708 


1798 


3274 


0998 


3119 


34 


00630 


1601 


2914 


0888 


2776 


35 


00561 


1426 


2595 


0791 


2472 


36 


00500 


1270 


2312 


0705 


2203 


37 


00445 


1131 


2058 


0627 


1361 


38 


00397 


1007 


1836 


0560 


1749 


39 


00353 


0897 


1633 


0498 


1555 


40 


00314 


0799 


1452 


0443 


1383 



Eshbach, 0. W , op. cit., p. 1-150. 

6 Assumed specific gravities or densities in grams per cubic centimeter- copper, 8.89; aluminum, 
2.71; brass, 8.47. 



408 THE USES OF COPPER 

nesses, and weights per square foot for the Brown and Sharpe 
series. 

Copper sheet is used for producing fabricated objects by cold- work- 
ing stamping, spinning, etc.; also by welding, soldering, or brazing. 
The sheet may be used directly for roofing, sheathing of boats, flash- 
ings, drains, and numerous other purposes. 

Copper is often used as a refractory material, for such purposes as 
tuyere jackets, furnace doors, and locomotive firebox plates. Copper 
is not particularly resistant to high temperatures, but its high thermal 
conductivity makes it valuable as a refractory the heat is conducted 
through the copper so rapidly that its temperature never rises danger- 
ously high. When copper is used in this way it must be cooled by 
circulating air or water. For moderately high temperatures when it 
is desired to retain the hardness of the copper, arsenic- or silver-bearing 
copper is used. For high-temperature service (furnace doors, for 
example) where the copper is exposed to reducing gases, oxygen-free 
copper must be used to avoid embrittlement. 

TUBING AND PIPE 

Copper pipe can be made by rolling copper sheet into a cylinder and 
welding the seam, or seamless tubing can be made by piercing billets 
and rolling them over a mandrel to give the proper size bore and wall 
thickness. Small tubes are commonly made by the piercing method; 
larger pipes are made from sheet copper. 

Copper tubing is widely used in radiators, refrigerators, air con- 
ditioning equipment, and similar equipment where maximum heat 
transfer is desired. 

SUMMARY 

It has not been possible to enumerate all the detailed uses of copper, 
and we have indicated only the important general classifications. 
Practically all pure copper used commercially is in one of the three 
forms mentioned, with distribution about as follows: 1 

Per Cent 

Wire and rod 58 

Sheet and strip 29 

Tube 13 

All the commercial uses we have mentioned, it will be noted, involve 
mechanically formed shapes. Pure copper does not make satisfactory 

1 Stoughton, B., and Butts, A., Engineering Metallurgy, 3d ed., p. 308, McGraw- 
Hill Book Co., New York, 1938. 



ELECTROPLATING AND ELECTROFORMING 409 

castings and cannot be used to produce finished objects by casting. 
Brass, bronze, and other copper alloys can be used for castings, but 
not pure copper. 

ELECTROPLATING AND ELECTROFORMING 2 

The object of electroplating is to deposit a layer of copper on the 
surface of another metal or alloy by electrolysis, and the purpose is to 
secure a composite object that will have the physical properties (or low 
cost) of the underlying metal or alloy plus the surface properties of 
pure copper. Electroformmg is a process of making an electrodeposited 
shell to form a " negative " of some object. 

Electroplating. In general the electroplating of copper resembles the 
electrolytic processes that we have already discussed. The object to 
be plated is immersed in the electrolyte and made the cathode; anodes 
may be either soluble or insoluble. 

In electroplating, the deposit must be smooth and coherent, and this 
is best obtained when the deposit consists of very fine crystals. Thick 
deposits invariably become roughened because the crystals become 
larger as they grow away from the cathode (Fig. 8, p. 268) . A number 
of factors influence the crystal size and coherence of the cathode deposit 

1. Current density. 

2. Concentration of metal ions in the electrolyte. 

3. Concentration of other salts in the electrolyte. 

4. " Throwing power " of the electrolyte. 

5. Hydrogen-ion concentration. 

6. Use of addition agents such as glue. 

7. Nature of the base metal. 

8. Conductivity of the solution. 

The " throwing power " of an electrolyte refers to its ability to de- 
posit an even layer of metal in holes and crevices of irregular objects; 
most conducting salts and substances that increase cathode polarization 
tend to improve throwing power. 

In general the proper composition of electrolyte, current density, etc., 
is determined by experience in dealing with different metals. For 
copper plating, there are two types of " plating baths " or electrolytes 
in common use: 

1. Acid Sulfate Bath. This bath consists of an aqueous solution 
of CuS0 4 and H 2 S0 4 , and the concentration of the two may vary 
between wide limits. The weight of bluestone (Cv30 4 -5H 2 0) added 

2 Mantell, C. L., Industrial Electrochemistry, 2d ed., p. 187, McGraw-Hill Book 
Co., New York, 1940. 



410 THE USES OF COPPER 

is usually held between 150 and 240 grams per liter. Temperature of 
the acid sulfate bath is maintained at 25 to 50 C, and a cathode 
density of 15 to 40 amperes per square foot is used. 

2. Cyanide Bath. This is essentially a solution of the complex 
cyanide of sodium and copper which probably has the formula 
Na 2 Cu(CN) 3 , plus an excess of sodium cyanide. In some cases 
Rochelle salts (potassium sodium tartrate) are used in connection with 
the cyanide. The copper content of these solutions will range from 
22 to 26 grams per liter, temperature from 35 to 40 C, and cathode 
current density from 3 to 14 amperes per square foot. 

Plating baths having low concentrations of the metal ions are best 
for securing fine-grained deposits which are easily polished, and the 
low ionic concentration is produced not by using dilute solutions but 
(1) by the use of salts showing low ionization or by the addition of 
another salt having a common ion to depress the ionization of the 
metallic salt, or (2) by using a compound in which the metal ions 
are produced by secondary ionization. Sodium copper cyanide is an 
example of the latter; in water the principal ionization is of the type 

Na 2 Cu(CN) 3 ^ 2Na+ + Cu(CN) 3 ~~ 

so that the bulk of the copper is tied up in the complex negative ion. 
There is a certain amount of secondary ionization of the type: 

Cu(CN) 3 " ^ CuCN + 2CN" 
CuCN ^ Cu+ + CN" 

which produces the copper ions. 

Cyanide baths are used for the formation of thin platings and as a 
foundation layer for thicker coats. Thick layers are generally plated 
from the acid sulfate bath. 

Copper plating is widely used both to give a finished surface and as 
a foundation for the plating of other metals such as nickel. So-called 
" chrome plate " usually consists of a copper deposit followed by a 
nickel deposit with a thin deposit (" flash plating ") of chromium over 
the layer of nickel. 

Brass Plating. Brass (copper-zinc alloy) is the only alloy that is 
used for plating in commercial work on a large scale. It is plated 
from a solution of the double cyanides of copper and zinc, using a 
brass anode. Copper and zinc will precipitate together from such a 
solution because their deposition potentials are almost the same. In 
sulfate solutions, however, the deposition potentials are very far apart, 
and no zinc will deposit as long as there are copper ions in the solution. 



ELECTROFORM1NG 



411 



Electrof orming. Electrolytically deposited copper can be deposited 
on an object to form a shell which is a " negative " replica of the 
object. This method is used for reproducing a printer's set-up of type 
(electrotyping) , engravings, and medals; for the reproduction of phono- 
graph matrices ; and for the manufacture of seamless tubes and sheets 



Copper Reel 




(Shakespeare, Am. Insl Mm. and Met. Eng. Trans., Vol. 106, p 442, 1933) 

FIG. 1. Sectional Elevation of First Stage Machine Used in the Manufacture of 

Electro-Sheet Copper. 

by electrodeposition. When it is desired to plate metal on a non- 
conducting object such as a wax phonograph record the surface is 
made conducting by giving it a light coat of graphite. 

Seamless tubes are made by depositing a layer of copper on a rotating 
mandrel which serves as the cathode; the mandrel is coated with 
material which facilitates removal of the finished tube. 

Very thin copper sheets have been formed by electrodeposition on 
a belt moving continuously through the solution, the product being 
taken off the belt where it passes out of the solution. A process 
developed at the Raritan copper plant 3 utilizes two stages (Figs. 1 
and 2). In the first stage (Fig. 1) a slowlv rotating lead-covered 
copper drum serves as the cathode, and insoluole lead anodes are used. 

8 Shakespeare, William M., Anaconda Electro-Sheet Copper: Am. Inst. Min. & 
Met. Eng. Trans., Vol. 106, p. 441, 1933. 



412 



THE USES OF COPPER 



The sheet produced measures 0.00135 inch thick and weighs about 
1 ounce per square foot. This 1 -ounce sheet is then sent through 
the second stage (Fig. 2), where it can be built up to any desired 
thickness. 



Starling Shi 



Steam Roll 
Wash Rolls X^ ^Finished She 

,@> 




Sectional Elevation 
(Shakespeare, Am Inst Mm. and Met Eng Trans., Vol 106, p 443, 1933) 

FIG. 2. Sectional Elevation of Second Stage Machine for the Manufacture of 

Electro-Sheet Copper. 

In the second stage the sheet is conducted through a series of 
depending loops spaced between standard refinery anodes. The sheet 
material produced is sound and uniform, both m physical properties 
and in gauge. 

COPPER POWDER 4 - 5 

In recent years the use of powdered metals has become of con- 
siderable importance, and copper is included among the metals which 
are being produced commercially in the powdered form. Copper pow- 
der is a finely divided powder of pure metallic copper, its purity de- 
pending on the copper from which it is made. 

Powdered copper is used to some extent as a pigment, but the most 
important use is in the manufacture of solid metals and alloys by 
various processes for compacting the powder into solid pieces. The 
technique of powder metallurgy has made it possible to produce solid 
compacts or " alloys " which could not be made by the standard method 

4 Noel, D. , Shaw, J. D., and Gebert, E. B., Production and Some Testing 
Methods of Metal Powders: Am. Inst. Min. & Met. Eng. Tech. Paper 928 (Metals 
Technology), June 1938. 

5 Goetzel, C. G., Powder Metallurgy of Copper: Metals and Alloys, Vol. 12, Nos. 
1 and 2, pp. 30, 154, 1940. 



MANUFACTURE OF COPPER POWDER 413 

of melting and casting. Thus it is possible to make solid copper- 
graphite compacts by this method, but copper-carbon alloys cannot 
be made because carbon is insoluble in molten copper. Compacted 
alloys can also be made which have controlled porosity " self- 
oiling " bronze bearings are an example. These are made by com- 
pacting copper and tin powders cold, and then sintering the section. 
Tin diffuses into the copper to form grains of alpha bronze, and between 
the grains are left the capillary pores through which the lubricating 
oil can flow. 

Manufacture of Copper Powder. There are a number of methods 
for preparing metal and alloy powders; the following are applicable 
to copper. 

1. Machining. Turnings, cuttings, filings, etc., produce a relatively 
coarse irregular powder which is useful for certain purposes. There 
is not much pure copper powder produced in this way because copper 
is too soft and tough to machine readily. 

2. Milling. Copper and other ductile metals can be ground to 
powder in special ball mills or stamp mills. The metal is fed into the 
mill as small pieces of thin sheet, and a lubricant is used which prevents 
the particles from " welding " together. Powder produced in this way 
is flaky and has a low apparent density ; it is especially suited for pig- 
ments because of the " leafing " characteristic of the powder flakes 
which give a good covering power. Copper flake is used extensively in 
the manufacture of motor brushes for commutator and collector rings, 
being compounded with carbon or graphite and occasionally smaller 
amounts of lead or tin. The flake type of copper powder has an ad- 
vantage for this work owing to its tendency to laminate during molding. 

3. Reduction of Copper Oxide. Copper powder can be produced 
by treating copper oxide powder in a reducing atmosphere at high 
temperatures (but below the melting point of either metal or oxide). 
Powders reduced from oxides have a granular spherical shape and a 
spongelike structure that makes them particularly adapted for molding 
work. Copper oxide scale formed in hot-rolling copper is the raw 
material commonly used for this method. 

4. Chemical Precipitation. The "cement copper" pioduced by 
precipitating copper from its solutions is a type of copper powder, but 
it is generally too impure to be used directly. 

5. Electrolytic Deposition. A large amount of copper powder is 
produced electrolytically, using essentially the same method as that 
employed in copper refining. Special conditions in the electrolyte 
cause the deposit to form as loosely adherent fine crystals which can 
be removed by scraping or tapping the cathode. The electrolyte con- 



414 THE USES OF COPPER 

tains less copper and more acid than refinery electrolytes, and a higher 
current density is employed. Hydrogen evolution at the cathode, or 
the addition of certain colloidal materials to the electrolyte aids in the 
production of fine powders. 6 Electrolytic powder grains have a " fern- 
like " appearance under the microscope, a structure well adapted for 
molding work. The grain size of the particles is controlled by the 
electrolyte composition and current density used. 

COPPER COMPOUNDS 

A small amount of copper is consumed in the manufacture of 
chemicals, insecticides, preservatives, etc. Considerable quantities 
of "blue vitriol" or "bluestone" (CuS0 4 -5H 2 0) are produced as a 
byproduct of electrolytic plants. 

THE ALLOYS OP COPPER 

Most of the " metals " used commercially are really alloys, and 
there probably would be some justification in considering some of the 
grades of commercial copper as alloys tough-pitch copper as a 
copper-oxygen alloy; high-resistance Lake copper as an alloy of 
copper, silver, and arsenic. Certainly the elements present other than 
copper have a marked effect on the properties of the metal. 

Copper is widely used in the unalloyed form for the following 
reasons: 

1. Its high electrical and thermal conductivity makes it invaluable 
for many uses. Alloying invariably lowers both the electrical and 
thermal conductivity. 

2. Its ductility makes it useful in many operations where severe 
cold-working is necessary. 

3. It resists corrosion by certain corrosive agents; other materials, 
however, corrode copper quite rapidly. 

Copper alloys are used because they are more satisfactory than 
cbpper for certain purposes, among which are the following: 

1. Most copper alloys are harder and stronger than copper. Copper 
is rather soft and ductile in the annealed state as compared with 
annealed brass, bronze, or other copper alloys. Many of these alloys 
can be work-hardened to produce a much harder and stronger 
material than cold-worked copper. 

2. Some copper alloys are used to make castings brass and 
bronze valves, for example which cannot be formed by any other 
method. Copper cannot be used to make satisfactory castings. 

6 Mantell, C. L., op. cit., p. 219. 



COPPER-BASE ALLOYS 415 

3. Copper alloys are used to make objects which must be machined 
cut, threaded, milled, etc. Copper is soft and tough and difficult 
to machine satisfactorily. 

4. Many copper alloys show resistance to corrosion equal or superior 
to that of copper. 

5. Alloys such as brass (copper-zinc) which contain a cheaper metal 
than copper, are equally satisfactory for many uses, and are less ex- 
pensive. 

6. Certain copper alloys can be hardened by htat-treating processes 
(notably the copper-beryllium alloys), but pure copper does not 
respond to this treatment. 

7. Some copper alloys display marked elasticity ; this is almost totally 
lacking in copper. 

We shall divide the general subject into two parts: (1) copper-base 
alloys in which copper is the predominant metal, and (2) other 
alloys containing copper. 

COPPER-BASE ALLOYS 

Brass and Bronze. Brasses and bronzes are the most important of 
the copper-base alloys. Originally the term brass was used to 
designate a copper-zinc alloy, and bronze for the copper-tin alloys, but 
commercial usage has modified this terminology, as may be seen from 
Tables 7 and 8. Many of the brasses and bronzes contain three or 
four alloying elements, and there are " bronzes " which contain no tin 
at all. The tendency is to call the reddish-colored alloys bronzes 
regardless of their chemical composition. 

Constitutional diagrams of the copper-zinc and copper-tin alloy 
systems are given in Figure 3. Most commercial brasses contain less 
than 38 per cent zinc and hence consist entirely of the alpha solid 
solution of zinc in copper, or " alpha brass." A few alloys, such as 
Muntz metal, contain more than 38 per cent zinc, and these alloys 
are made up of alpha and beta brass. The beta constituent is brittle 
at room temperature and these alloys cannot be cold-worked. There 
are few commercial alloys which contain any of the other constituents 
shown on the diagram these are all very brittle, and alloys con- 
taining them can be shaped only by casting. 

Alpha brass consists of homogeneous grains of the alpha solution of 
zinc in copper. It ranges in color from a copper red to a bright yellow. 
Brass containing relatively small amounts of zinc (5 to 20 per cent) is 
known by various names red brass, low brass, commercial bronze, 
etc. Brass containing 30 to 40 per cent zinc is known as common 



416 



THE USES OF COPPER 




400 





10 20 30 40 50 60 70M80/90'100 
Per Cent Zinc 7+(Y+O)+n ^ 
e+(7 + O 




Copper-Zinc 



10 20 30 40 
Per Cent Ttn-Weight 

Copper-Tin 



1600 




20 40 60 80 
Percent Nickel By Weight 

Copper-Nickel 



2 4 6 8 10 12 
Per Cent Beryllium, By Weight 

Copper-Beryllium 



14 





10 20 30 40 50 60 70 80 90 100 
fer Cent Copper, By Weight 

Copper-Silver 

(Prom Metals Handbook, Am. Society for Metals, 1936 Ed.) 
FIG. 3. Equilibrium Diagrams for Some Copper Alloys. 



600 



"0 2 4 6 8 10 12 14 16 18 
Per Cent Aluminum 

Copper-Aluminum 



BRASS AND BRONZE 417 

yellow brass, high brass, cartridge brass, etc. In addition to copper 
and zinc, many brasses contain other metals (Tables 7, 8) , added for 
various reasons. Lead is added to brass in small quantities to improve 
its machining qualities. Tin is added to some brasses to improve their 
mechanical properties. 

Brass is an alloy that is cheaper, stronger, and harder than copper; 
it is less ductile but still retains excellent formability. It can be 
worked hot or cold and is made into sheet, tubes, pipe, stampings, and 
wire. The corrosion resistance of brass is comparable with that of 
copper for most uses. Brass is widely used for castings gears, 
propellers, steam and pipe-fittings a purpose for which copper cannot 
be used. 

The tensile strength of brass increases with the zinc content, and 
annealed yellow brass will have a tensile strength of about 45,000 
pounds per square inch (annealed copper = 32,500 psi) . Cold-working 
may raise the tensile strength to 120,000 pounds per square inch 
(cold-worked copper = about 60,000 psi). 

Brass can be softened by annealing after cold-working but cannot 
be hardened by heat-treatment. Rolled brass is subject to " season 
cracking," which is thought to be due to a combination of corrosion plus 
a readjustment of the strained crystals. Greater care in rolling and 
annealing will make season cracking less likely to occur. 

Brasses for special purposes may contain iron, manganese, and 
aluminum, as well as the more common addition agents, lead and 
tin. The data in Tables 7 and 8 give the composition and properties of 
the most important of these alloys together with notes as to their 
common uses. 

Bronze (Fig. 3) is a copper-tin alloy, and most commercial bronze 
consists entirely of the alpha solid solution (alpha bronze). True 
bronze is superior to brass in many respects, but tin is much more 
costly than zinc so bronze is not ordinarily used where brass will do. 
The recognized superiority of bronze probably explains the tendency 
to call copper-zinc alloys " bronze " when they have the characteristic 
reddish color that is associated with bronze. As in the case of brass, 
there are special bronzes which contain elements other than tin 
(Tables 7 and 8). Phosphor bronze is one of these special bronzes. 
Aluminum bronzes contain aluminum instead of tin. 

Both brass and bronze are fairly good electrical conductors but far 
inferior to pure copper in this respect. Bronze is superior to brass in 
mechanical strength and corrosion resistance, and it makes sounder 
and denser castings. 



418 THE USES OF COPPER 

Copper-Nickel Alloys. Copper and nickel are soluble in all propor- 
tions in the solid state (Fig. 3) and zinc can be dissolved in this solid 
solution. The cupro-nickels contain copper and nickel only; " nickel 
silver " is a ternary alloy of copper, nickel, and zinc (Table 8) . In 
general these are white alloys and have the same general properties as 
brass and bronze; they are much more corrosion-resistant, however. 

Beryllium Copper. Beryllium copper has come into prominence in 
the last 10 years or so, and is an example of an alloy which can be 
hardened by heat-treatment. Commercial beryllium copper contains 
about 2.25 per cent Be, and Figure 3 shows that the normal alloy 
consists of the alpha and gamma phases; the gamma phase is a hard 
compound of copper and beryllium. When the alloy is heated, the 
gamma phase goes back into solution, and if the alloy is quenched the 
gamma phase is retained in a supersaturated alpha solution. Heating 
the quenched alloy then causes the hard gamma phase to precipitate 
throughout the alpha grains and key the slip planes so that the alloy 
becomes much harder and stronger. This is the type of heat-treatment 
known as a precipitation hardening (age-hardening, if the precipitation 
in the quenched alloy takes place at room temperature) . 

OTHER COPPER ALLOYS 

Copper is used in many other alloys strong aluminum alloys, 
magnesium alloys, Monel metal, and copper-bearing steel. 

Monel Metal. Monel metal is a widely used alloy containing about 
two-thirds nickel and one-third copper. It is about the equal of mild 
steel in mechanical properties, and is especially resistant to corrosion. 
Monel metal is a natural alloy; i.e., it is made by reducing copper-nickel 
mattes in which the nickel and copper ratio is about the same as in 
the alloy. 

Copper-Bearing Steel. Steel for sheet and tubes is being manufac- 
tured which contains about 0.30 per cent copper. A considerable 
quantity is consumed for this purpose in spite of the small percentage 
of copper used in the alloy (Table 1). The principal advantage of 
copper-bearing steel is its increased resistance to atmospheric corrosion. 



COPPER-BASE ALLOYS 



419 









Ji 




RI5 

5 












> $ >> 




i 




8 yj 




*^ o 

l 




5 




o 3 




>,^ 




1 




w ? 




T 2 
8 5 




.3 
3 




Deoxidized with 5 
free Mn. Electi 




Aluminum-bronz 
should not be s 
a high tempera! 




Ills 












it! 

" 
















I a ^ H o ^ . 










00 


g 


Jj!*3t!! 











h 


a 


i g ^-.o g- 

G a ^o S, M 






jjj 
PH 





b- W 


||. 2 !V 


g 


S 


Q 


IN, O <N 


i 

w g 

1-3 s 


K 


a 

i 




1 
N 

o 
Pi 

H 






i d 

2 a d x? 

sI - 


fS 


>oJt5? 


1 





0) ** 

a d 




<Q 


I 


r.^^ 


I 


1 


g ? SM 




^ 


h 


^^ 


OQ 




2 | 2^ 







w 


r^^oo 

vO " 
IS (N 






V 














fe 














d 












i 
f 


B 










d 








1 


N 














=3 








* 





Ox 0* 


; 












a 




1 




S, 




h 




| 




g 1 




ll 








( 




^ 





Aluminum-bronzes. These alloys 
should not be slowly cooled from 
a high temperature if they are to 


i3 $ 

i|- 
a ? ~ 
s |w 

Iji 3 

! 9l ? 

Ul 

Sla 


and other equipment for chemical 
industry. 


Everdur. Resistant to H 2 SO4, HCI 
(in absence of air), sea water, 
caustic soda, phenol, etc 


Silicon-brass, for hardness. 


Manganese-bronze. For strength, 
toughness, and hardness. 


Manganese-bronze. Propellers, en- 
gine frames, parts requiring 
strength and toughness. Resist- 
ant to sea water and acids. 










fN 


o 




: 







Is 


o 


s 


130-150 


150-180 


o 

I 







I 




: 










PH 






fS 


00 

5 









5 


Q 

g 


IN, O <N 


(S 


? st 


JN 




<N 




S5 
















Pi 
H 




s 


,AS 

rs 




I 






1 

M 


r.^^ 


5? 


o 
r> 


I 


5 


85-100 


5 


O 

W 


r^3 

vO * 
IS (N 


O 


I? 







1 


I 






- 


n 






(A 


h* 











- 







5 










l/N 



















: 


N 


s 




* 





0- 






- 






S 


00 


00 
00 





00 


oo 


00 




Aluminum 
bronze 


Aluminum 
bronze 


Aluminum 
bronze 


Everdur 


Silicon 
bras* 


Manganese 
bronze 


ji 



8 
1 



420 



THE USES OF COPPER 



I 




I 

I 



ffl 



C 

O 



u 



fc 





o. 
O 



g 

5 1 






I 





> < 
-3 





J5 

n 

a 

jo 

"3 ^ 
t S 

W) 

as g 

II 

m 


Bronze for water-tight castings, 
underwater fittings, machine 
parts Non-corrosive 


Gear bronze Very resistant to 
abrasion For heavy-duty gearq 
and worm wheels. 


Steam fittings Machines well. 


S, 

bO 

"S fl 
(U 

f! 

S ? 

> >-< 


Red brass for pump bodies, valves, 
steam fittings, beanng backs, and 
metal patterns. 





















iA 


IA 


lA 


S 


o 


















J> 


lA 


iA 


J> 


iA 


J, 




"*" 


10 


NO 


tA 


OO 


iA 




I* 




!A 






S 




fA 




rvi 










i= 




NO 




rs 


i- 




lA 
CA 

1 


vA 


CO 

1 

NO 







o 







O 


lA 


O 


eo 


o 




1 


A 


1 

iA 


.A 


1 


i 




O 


>A 


iA 

fA 


^ 


IA 


1 







S 


O 


^ 


^ 






OO 

I 


VA 

rsi 







2 


I 






r-i 


*"~ 




CNl 




P- 


o 











S 

CJ 









O 




o 


H 




* 




lA 












a 


rA 


^ 










N 














^ 


r-i 











A 


PH 















C 


OO 


00 


- 


00 


2 


,n 





CO 


OO 
00 


o 


CO 


CO 


A 
00 








a 




o> 


2 




a 


g 


^ o 


c 


d 






o 


o 


* - 





2 


*5 




PQ 


PQ 


O 


Q 




w 



COPPER-BASE ALLOYS 



421 



" O 


OJ 


" ^ 


-d >> 


^ +- 


c 




s ? 


g -d 


o 


2 "ii 





Semi-red brass Very resist 
atmosphenc corrosion For 
head electrical fittings ai 
oil and water pumps 


For hardness Difficult to ms 


Red brass For low-pressure 
and fittings (Radiator v 
Machines easily 


&l 

t-c ^^ 



Ij 

Ml 



















o 


sO 




oo 


sO 










A 


J, 


JL 


vA 




vA 









fN 






o 

I 









fN 







ff\ 

fN 




1 



ro 

fN 


VA 
fN 





vA 


O 

2 


SI 


* 


O 
fN 







fN 





































O 


sO 










* 





fN 
fN 


VA 





vA 

00 


00 





S 


^ S 





1 

Ja 


1l 


JP 


1=3 
ft) 


"8 


Si 






T) 


i, t 


^ 

















OJ 








c 


For oil feed c>lmder top cover 
for automobiles 


For light castings, builders' 
ware, ornamental work 


For common castings where c 
ness and good machining pi 
ties are main considerations 


"Na\al brass" Miscella 
work for strength 










o 


o 


o 




sO 

r 


sO 


< 














, 






VA 


VA 








i 


vA 


rs 


o 


vA 


O 

f 

A 


vA 
CO 


o 

1 

vA 


fN 




fN 




00 


o 


vA 


O 


fN 


i 


1 

O 


^ 


- 



fN 


vA 


o 

vA 




~~ 




fN 




























fN 


O 


ro 


^ 















- 






- 


- 


- 








fN 


^ 


sO 


sO 


sO 


. B 

|| 


|| 


.2 2 

15 ^ 


1 

'd 



>> a 

III 


ti 1 ! 


I 


stant to salt 
bleaching so 
rchltectural t 


1 Resistant 
rrosion For 
n, bathroom 
gs, builders' 


4) r-l 

-* 

*i 


Ifs 


o t> 

fi 


l?i 


o^3 


g "3 e 


6s,,; 


08 ^ aj 


.g a a J! , 


2 10 


i "S 




iS N 1 


s55 


S -2 J2 ^ 


1 o g 


-fj 


ffl 


c2 






8 si 






I 






8 ? 






c "o - 






C J 






| 3 ^ 






|^' 






c 2 






s ;M 






00 K JS 






TH Cl W 






rt "" " 


o 





ill 


VA 


iA 


c ^ 


I 




sS 


VA 
CN 





i r 


fN 


" 


p o tj 






S ^- s 

1 





vA 





fN 


O ofi O 


VA 





| 


^ 


o 


o, E 

O o 



VA 


vA 
<N 


gii 

;E 






M i 


VA 




O" 3 C3 


VA 
sO 


O 

vA 
vA 


on -^ 


O 


-JJ 


H ^ w 




2 


^83 


S 


H 




^ 


CQ 


o 



422 



THE USES OF COPPER 



TABLE 8. CHEMICAL AND PHYSICAL PROPERTIES 

(Variations must be 



Material 





Approximate Composition, 
per cent 


Tensile 
Strength, 
Ib per sq m 


Elonga- 
tion in 
2 in, 
per cent 


Yield Point, (g) 
Ib per sq m 


Johnson's 
Elastic Limit, 
Ib per sq in 


Modulus of Elasticity, Ib I 
persqinXlO-s.Hard*] 


^S, 

d 


5 


1 


c 
H 




o 

1 


1 


~o 

S 


i 


"O 

A 


1 


i 

w 


I 


Copper 


S 

w 

R 


99 90+ 
99 90+ 
99 90+ 








51,000 
60,000 
50,000 


32,500 
38,000 
32,000 


4 
3? 
18 


37 
36c 
38 


48,000 
39,000 
46,000 


12,000 
15,000 


36,500 
35,000 


5,000 


16 


Deoxidized copper 


T 


99 90+ 








Phos- 
phorus 
present 


50,000 


35,000 


10 


35 


48,000 












S 
R 
W 


99 90+ 
99 90+ 
99 90+ 








present 
present 
present 


55,000 
58,000 
60,000 


35,000 
35,000 
35,000 


5 
5 

26f 


35 
38 
35 


44,000 


15,000 
20,000 






16 


Primer gilding 


S 


97 00 


3 00 








55,000 


35,000 


5 


37 


38,000 








Commercial bronze 
95% 


S 


95 00 


5 00 








55,000 


35,000 


5 


38 


39,000 


11,000 






15 


Commercial bronze 
90% 


s 


9000 


10 00 








67,000 


37,000 


3 


40 


53,000 


11,000 






15 


Red brass 85% 


s 

T 


85 00 
85 00 


15 00 
15 00 










75,000 
68,000 


42,000 
42,000 


4 
6 


43 

42 


71,000 
64,000 


18,000 
19,000 


49,000 




15 


Red brass 80% 


S 


80 00 


20 00 




85,000 


43,000 


4 


50 










150 




\\ 


80 00 


20 00 








125,000 


49,000 


2e 


43e 












Brazing brass 


s 


75 00 


25 00 








80,000 


47,000 


5 


45 












Spring brass 


s 


72 00 


2800 








76,000 


47,000 


4 


55 


38,000 








14 


Cartridge brass 


s 


70 00 


30 00 








86,000 


45,000 


4 


50 












Cartridge brass 
Eyelet brass 


s 
s 


69 00 
68 00 


31 00 
32 00 








85,000 
78,000 


46,000 
46,000 


4 

5 


58 
58 


55,000 










Drawing or spin- 
ning brass 


s 


66 67 


33 33 








76,000 


46,000 


5 


52 












Commercial high 
brass 


s 

R 


65 00 
65 00 


35 00 
35 00 

3700 
37 00 
37 00 


_ 







76,000 
70,000 


45,000 
45,000 


5 

15 


60 
50 




12,500 


^ 48,000 


7,500 


14 


Brass wire 


R 

S 

w 


63 00 
63 00 
63 00 




70,000 
84,000 
125,000 


50,000 
48,000 
50,000 


12 
4 
2f 


50 
50 

50r 










140 



* For some alloys the figures given are for a 
temper slightly different from that commonly 
known aa "Hard" 

t Compared to water at 4 deg cent. 
JSoft 
R Rod 
8 Sheet 
T Tube 
W Wire 



a Temper not known 

6 Determination 

r Circular 73, U S Bureau of Standards 

d Scientific Paper 410, U S Bureau of 

Standards 

e Elongation of wire, per cent m 10 m. 
/ Corning Glass Works 
g Yield point taken as the load producing an 

extension under stress of 75 per cent* 



COPPER-BASE ALLOYS 



423 



OF VARIOUS WROUGHT COPPER-BASE ALLOYS 

expected in practice) 





Brmell 


Rockwell 




+. 









Electrical 








Hard- 


Hard- 




s 




..5 




Propf rti-s 






Shearing 


nessNo 


ne^s No 


Bend 


bD 




9 


s 


at 20 C 


^ 




Strength, 




"B" 


Test, 


0) 




3 










Ib per sq m. 


BalT 
500-kg 


1/16-m 
Ball, 


dog 


1 






w 


5 


^ 





Uses Remarks 




Load 


100kg 







C 


~ 


a -" 


sS 


-c g w 


tj 




^ 




, 




w 




^ 




a 


*2 


z? 


32 


1 i 


P w O 


3 "^ 




"O 




~CI 




T3 




-a 




43 








e v 




73 i- * 


1 




1 


s 


1 


& 


A 


3 


A 


r?. 


S 


1 


& 


gx 

CJ 


1 


1 M 


^U 






21,000 


103 


42 


58 


Too 


180 


180 


1083c 


8 89 


3217 


177 


10 371 


100 


9225 


For electncal parts requiring 












Soft 




















high conductivity, also for 














180 


180 
















rods, sheets, tubes and wire 














180 


180 
















for commercial purposes 










58 


Too 


180 


180 


10836 


8 936 


323 


177 








Copper pipe and tube Tougher 












boft 




















than electrolytic copper and 










61 




180 


180 
















has better corrosion resist- 














180 


180 
















ance Welding rod. May 














180 


180 
















also be deoxidized with sili- 
































con, manganese, calcium bo- 
































nde, and other deoxidizers 






107 


42 


65 


Too 


180 


180 


1075* 


8 890d 


321 


177 


15 62 


66 40 


59 


Primer gilding 












Soft 


























110 


43 


68 




180 


180 


I065x 


8 866d 


320 


181 


18 98 


54 6 


576 


For jewelry trade and manu- 
































facturing where soft, pliablo 
































metal is required 




25,000 


115 


50 


75 


1 


180 


180 


1045x 


8 804d 


318 


182 


25 36 


40 90 


446 


For window screen wire and 
































automobile radiators on ac- 
































count of resistance to corro- 
































sion and atmospheric action. 






135 


52 


82 


10 


180 


180 


1020z 


8 7456 


316 


187 


28 03 


37 


38 


For architectural work on ac- 
































count of color and resistance 
































to corrosion, also for hard- 
































ware and pipe 


43.000R 


27,000 


150 


53 


86 


11 


180 


180 


1000* 


8 667d 


313 


191 


31 95t 


32 5J 


335 


For its color and resistance to 
































corrosion and atmospheric 
































action 














180 


180 










36 90y 


28 ly 




Fourdnnicr wire 






157 


53 


87 






180 


980i 


8 594d 


310 


196 


34 6t 


30 OJ 


31 


For parts or articles to be 
































brazed or silver soldered, etc. 






158 


53 


88 


20 




180 


965x 


8 553d 


309 


198 


36 3 


28 60 


295 


For turbine blades, best spring 
































stock, brass rod, etc 


% 




157 


53 


87 






160 


955x 


8 528d 


308 


199 


37 61 


27 58 


290 


For primers, shot shells, car- 
































tridges, seamless tubes, etc. 






156 


53 


87 


22 




180 


950x 








37 55 


27 60 


290 


For cartridges, etc. 






156 


53 


87 


22 




180 


945x 


8 506 


307 




37 95 


27 30 


289 


For eyelets, cartridges, drawn 
































shells, etc 




30,000 


153 


52 


86 


20 




180 


938j 


8 476d 


306 


201 


40 12 


25 85 


287 


For deep drawing, cartridges, 
































etc 






153 


52 


85 


30 




180 


930x 


8 460d 


306 


202 


38 68 


26 8 


285 


For a large variety of articles, 




23,000 




























lamp fixtures, automobile ra- 
































diat rs, and for ornamental 
































purposes Not good for ex- 
































posure to weather 
















180 


920j 


8 437d 


305 


205 


39 97 


25 95 


285 


For nvets. pins, screws, and 
































other heading operations. 



h Jenkins and Hanson constitution diagram 
j Average linear coefficient per cleg cent from 
25 to 300 dcg cent Tests on rod vScieii- 
tific Paper 410, U S Bureau of Standards 
k At 18 1 deg cent 
m Cold worked and heat-treated 
n Guertler-Tammann constitution diagram 
p Annealed, quenched, and heat-treated 



r Smith constitution diagram 

t Stockd Je constitution diagram 

u G-c-il, per BOO per eq cm per deg cent per 

cm at 20 deg cent 
v Tafel constitution diagram 
x Bauer and Hansen constitution diagram 
y Hard at 25 deg cent 
z Heyoock-Nevme constitution diagram 



424 



THE USES OF COPPER 



TABLE 8. CHEMICAL AND PHYSICAL PROPERTIES 

(Variations must be 



Material 


S 


Approximate Composition, 
per cent 


Tensile 
Strength, 
Ib per sq in 


Elonga- 
tion in 
2m, 
per cent 


Yield Point, (g) 
Ib per sq in 


Johnson's 
Elastic Limit, 
Ib per sq m 


Modulus of Elasticity, Ib 
persqin XlQ-e.Hard* 


1* 


G 


1 


a 




1 


I 


1 


1 


S 


I 


I 


I 


Muntz metal 


S 


60 00 


40 00 








80,000 


57,000 


9 5 


48 




20,000 






12 8 


Cap gilding 


S 


90 00 


9 60 


40 






65,000 


39,000 


4 


35 












Tube brass 


s 

T 


67 50 
67 50 


32 00 
32 00 


50 
50 






85,000 
50,000 


43,000 
44,000 


3 
5 


46 

45 




17,000 






4 Ot 


Butt brass 


S 


64 00 


35 00 


1 00 






80,000 


45,000 


5 


60 












Leaded commercial 
bronze 
Leaded red brass 
80% 


R 
R 


88 50 
78 50 


10 00 
20 00 

29 50 


1 50 
1 50 







60,000 
60,000 


35,000 
40,000 


3 
10 


30 
35 










5 


Leaded brass 


R 


69 00 


1 50 




65,000 


45,000 


10 


34 


33,000 








Clock brass 


S 


61 50 


37 00 


1 50 






80,000 


45,000 


4 


40 












Forging bra?s 


R 


60 00 


38 00 


2 00 






70,000 


50,000 


10 


45 


31,000 


22,000 








Free-cutting rod 
Oreide 


R 

S 


62 00 
87 25 


35 00 
11 50 


3 00 


1 25 




62,000 
80,000 


47,000 
45,000 


20 
4 


60 
40 


52,000 


32,000 






5 


Admiralty 


S 
W 
T 


70 00 
70 00 
70 00 


29 00 
29 00 
29 00 




1 00 
1 00 
1 00 




95,000 
125,000 


45,000 
57,000 
55,000 


5 

2e 


60 
42e 
60 












Naval brass 
Tobm bronze 


R 
R 

S 


60 00 
60 00 
60 00 


39 25 
39 25 
39 25 




75 
75 
75 




62,000 
75,000 
90,000 


54,000 
54,000 
54,000 


25 
25 

4 


40 
50 
40 


60,000 


25,000 
25,000 


72,500 


16,000 


50 


Fourdrmier wire 
Special bronze 
Signal bronze 


W 

s 
w 


81 00 
98 75 
98 25 


18 75 




25 
1 25 
1 75 




65,000 
100,000 


49,000 
40,000 
50,000 


4 
3e 


43e 
48 






73,000 






Phosphor bronze, A 


s 


96 00 






3 75 


P 
25 


90,000 


45,000 


4 


50 




18,300 


68,500 


11,100 


5 




s 

R 

S 


95 00 






5 00 




100,000 


50,000 


3 


55 


87,000 


23,000 


74,000 




5 


Leaded phosphor 
bronze, B 
Phosphor bronze, C 


94 00 
92 00 




1 00 


5 00 
8 00 




110,000 


50,000 
55,000 


3 


40 
70 


85,000 


20,000 
25,000 






4 


Phosphor bronze, D 


s 


89 50 






10 50 




115,000 


60,000 


5 


65 


95,000 


40,000 


91,000 






Free-cutting phos- 
phor bronze 
High-strength bronze 


R 
W 


88 00 
97 25 


4 00 


4 00 


4 00 
2 00 


Si 
75 


60,000 
120,000 


45,000 


20 


36 


50,000 








5 



* For some alloys the figures given are for a 
temper slightly different from that commonly 
known aa "Hard " 

t Compared to water at 4 deg cent 
JSoft 
R Rod 
S Sheet. 
T Tube 
W Wire. 



a Temper not known 

6 Determination 

c Circular 73, U S Bureau of Standards 

d Scientific Paper 410, U S Bureau of 

Standards 

e Elongation of wire, per cent in 10 in 
/ Corning Glass Works 
g Yield point taken as the load producing an 

extension under stress of 75 per cent. 



COPPER-BASE ALLOYS 



425 



OF VARIOUS WROUGHT COPPER-BASE ALLOYS Continued 
ezpeoted in practice) 



Shearing 
Strength, 
Ib per sq m 


Brine! 
Hard- 
nessNo 
10-mm 
Ball, 
500-kg 
Load 


Rockwel 
Hard- 
ness No 
"B" 
Vl6-m 
Ball, 
100 kg 


Bend 
Test, 
deg 


ting Point, deg cent 


Specific Gravity f 


3 
g 

& 

J.Q 


a 

<Q 


Coefficient of Expansion 1 
X 107 0) 


Electrical 
Properties 
at 20 C 


Thermal 
Conductivity (u) 


Usea Remarks 


Resistivity, 
ohms (mil, ft) 


Conductivity, 
per cent 
I ACS 


* 


i 


1 


S 


T3 




* 

T3 




w 





rt 


2 


S 






155 


80 


87 
90 


42 




180 


905-c 


8 396 


303 


208 


36 25 
24 63 


28 60 
42 10 


300 


For bolts, nuts, sheathing, pm 
wire, etc 
For caps and electncal parts 






158 


52 


15 








8 4956 


307 

319 
314 


38 68 


26 8 




Pipe and tube. 


35,000 




04 




85 
58 

87 


15 








8 830d 
8 698rf 


183 
192 


25 61y 
35 87j/ 


40 50j/ 
28 91y 


432 


For hinges, etc Also forthread- 
ing and similar operations 
P'or hardware, jewelry, etc. 
Free cutting 
For rivets, etc. Free cutting. 






42 
22 


54 
82 


13 
16 

55 




120 




8 562d 
8 44d 


309 
305 


200 


37 65 

39 10: 


27 55 
26 5t 


258 


For special shapes where high 
lead is detrimental to the 
bending or working of the 
stock. 
For clock and meter parts, 
pinions, and other articles 
where free milling is required. 
For hot forging. Machines 
easily. 


36,000 


28,000 


20 


54 


77 


20 


180 
180 


8856 


8 489d 


307 


204 


41 46 


25 


258 


For automatic machine work. 
Drills and turns easily. 












9356 


8 5356 


308 


202 

214 
211 


42 07 


24 65 


263 


For condenser tubes Resists 
action of sea water. 


37,000 
45,000 


33,000 
33,000 
33,000 


00 
65 


89 
90 


75 

93 




180 


8856 


8 4046 


304 


41 60 


24 93 


279 


For piston rods, propeller 
shafts, nuts, bolts, plates, 
etc Welding rod. 






75 
90 


71 

90 
96 


30 
30 





180 
180 
180 


075z 
070z 


8 7\2d 
8 89/i 
8 896 


315 
321 
321 


190 

178 


24 1 
29 62 


32 20 
43 
35 


341 
520 
350 


For Fourdnmcr wire 
For flexible metal hose 
For electrical purposes. 


54,000 


33,000 


60 
60 


180 
180 


050z 
050* 


8 886 
8 876 


321 
320 


82 18} 
56 46 


2 62} 
8 37 


150 

195 


For springs, electric switches, 
etc. 
For window weight chain. 
Bronze chain in general. 


60,000 
64,000 




200 
200 


70 
74 


99 
100 


75 
38 

52 




180 
180 

180 


1025z 
lOOOz 


8 9296 
8 8156 

8 786 


322 
318 

317 


182 
183 


56 46 
79 8 

98 10 


18 37 
13 00 

10 6 


199 
150 

121 


Phosphor bronze with good 
machining properties. 
For electric witches, contact 
fingers, diaphragms, radio 
parts, etc. 
For vciy stiff resilient springs 
(flat or coiled). Also for chain, 


35,000 




120 




75 






180 


10226 


8 866 


320 




86 43 


12.2 
12 


133 


For good machining proper- 
ties. 
For electncal purposes 



h Jenkins and Hanson constitution duigrara 

j Average linear coefficient per deg cent from 
25 to 300 deg cent Tests on rod Scien- 
tific Paper 410, U S. Bureau of Standards. 

fc At 18 1 deg cent, 
m Cold worked and heat-treated 

n Gucrtler-Tammann constitution diagram. 

p Annealed, quenched, and heat-treated. 



r Smith constitution diagram 

t Stockda'e constitution diagram 

u G-cal, per sec per sq cm per deg cent per 

cm at 20 deg cent 
v Tafel constitution diagram 
x Bauer and Hansen constitution diagram 
y Hard at 25 deg cent 
z Heycock-Neville constitution diagram. 



426 



THE USES OF COPPER 



TABLE 8. CHEMICAL AND PHYSICAL PROPERTIES 

(Variations must be 



Material 


Approximate Composition, 
per cent 


Tensile 
Strength, 
Ib per sq in 


Elonga- 
tion m 
2m, 
per cent 


Yirld Point, (s) 
Ib per sq m. 


Johnson's 
Elastic Limit, 
Ib per sq in 


k 

wx 














-39 

3 cr 


, 


a A, 


o 

a 


1 








* 
-o 


*, 


T3 

ft 


^ 


i 





1 


^ 


11 


i 


H U 


N! 


^ 








W 


S 


W 


& 


w 


$ 


W 


$ 




Super-nickel 1 


1 70 00 




30 00 










65,000 




30 












20% cupro-nickel S 


80 00 




20.00 








85,000 


50,000 


2 


30 












15% cupro-nickel S 


85 00 




15 00 








70,000 


45,000 


3 


30 


51,000 










30% nickel silver S 


47 00 


23 00 


30 00 








130,000 


72,000 


2 


35 












v 


* 47 00 


23 00 


30 00 








160,000 


75,000 


If 


35f 












Ambrac B S 


65 00 


5 00 


30 00 








105,000 


65,000 


2 


30 












R 


65 00 


5 00 


30 00 








85,000 


65,000 


10 


30 










20 


Y 


T 65 00 


5 00 


30 00 








130,000 


65,000 


2e 


30 












25% nickel silver S 


55 00 


20 00 


25 00 








110,000 


72,000 


4 


30 












Ambrac A S 


75 00 


5 00 


20 00 








85,000 


50,000 


5 


35 


77,000 


23,000 


57,000 


10,000 




R 


75 00 


5 00 


20 00 








80,000 


55,000 


10 


50 


70,000 


18,000 


52,000 


13,000 


19 OJ 


V 


T 75 00 


5 00 


20 00 








115,000 


55,000 


2e 


30e 












18% nickel silver S 


6500 


17 00 


18 00 








90,000 


58,000 


3 


40 


83,000 




72,000 




18 


18% nickel silver S 


5600 


26 00 


1800 








100,000 


60,000 


2 


40 












Y 


' 56 00 


2600 


18 00 








143,000 


60,000 


\e 


We 






103,000 




14.1 


15% nickel silver S 


64 00 


21 00 


15 00 








93,000 


58,000 


55 


40 






75,000 


22,000 




15% nickel silver 3 


5700 


2800 


15 00 








95,000 


55,000 


2 


35 




















Lead 
























Leaded nickel silver 3 


61 00 


25 00 


12 50 


1 50 






90,000 




5 














10% nickel silver 3 
5% nickel silver V 


65 00 
r 6300 


25 00 
32 00 


10 00 
5 00 








90,000 
135,000 


50,000 


3 

2e 


45 




11,000 






17 5J 










Al 
























5% aluminum S 


95 00 






5 00 






105,000 


52,000 


5 


70 












bronze 
































8% aluminum S 


92 00 






8 00 






120,000 


60,000 


4 


60 


60,000 








15 


bronze 
































R 


92 00 






8 00 






100,000 


60,000 


4 


60 






















Iron 






















8% aluminum J 


89 50 






8 00 


2 50 




125,000 


72,000 


5 


50 


80,000 


35,000 








bronze with iron 
































10% aluminum R 


90 00 






10 00 






125,000 


78,000 


5m 


36 


67,000 


41,000 








bronze 














7H 


















Aviahte R 


90 00 






9 50 


50 




88,000 




35 




43,000 











* For some alloys the figures eiven are for a 
temper slightly different from that commonly 
known as "hard " 

t Compared to water at 4 deg cent. 

i soft 

R Rod 
S Sheet 
T Tube 
W Wire 



a Temper not known 

6 Determination 

r Circular 73, U S Bureau of Standards 

d Scientific Paper 410, U S Bureau of 

Standards 

e Elongation of wire, per cent in 10 m 
/ Corning Glass Works 
g Yield point taken as the load producing an 

extension under strew of 75 per cent. 



COPPER-BASE ALLOYS 



427 



OF VARIOUS WROUGHT COPPER-BASE ALLOYS Continued 

expected in practice) 





Brinell 


ockwell 




^ 








Electrical 








lard- 


Hard- 




S 








Properties 






Shearing 


088 No 


ess No 


Bend 


bO 




- 




at 20 C 


, 








"B" 


Test 
















Ib per sq in 


Ball, 
500-kg 


/16-in 
Ball, 


deg 


a 


OJ 


K 


^^ 





/ 


r 


Uses Remarks 




Load 


100kg 




(X. 


O 




3 s "" 


f~ 


"*1 r 


| 




1 


^ 


% 


^ 


2 


^ 


3 


^ 


1 


$ 


1 


x 


II 









W 


$ 


W 





W 





a 





& 


f/j 


Q 




rf 


~> 


H 




















225n 


950 


323 


62/ 


18 4 


75 


069 


? or condenser turx s 










85 


7 5 




80 


200n 








60 2 


47 


087 


?or turbine blad* s and parts 
































where resistance to corrosion 
































and (rosiori is required 
















80 


175n 


956 


323 




27 


17 


112 


For bullet jackets 












61 




80 


140i> 


746 


316 




90 


58 




Has comparatively high elec- 
































trical resistance Used in 
































electrical instruments 






95 


63 


96 


32 




80 


2206 


8 866 


320 


62f 


231 8a 


4 47o 


068 


For resistance to corrosion and 


55,000 






























atmospheric action, also for 
































orrurnuital purposes 






208 


89 




60 




80 


135* 


8 726 


315 




259 Oa 


4 OOa 




For tableware, plated and uri- 
































plated 






160 


58 


88 


25 




80 


1506 


8 860 


320 


164 


172 


6 2 


092 


For resistance to corrosion and 


50,000 


33,000 




























atmospheric action, also for 
































ornamental purposes. 






170 


70 


91 


40 




180 


lHOu 


8 7526 


316 




175 


5 91 


080 


For silver-plated forks, spoons, 
































knives, hollow ware, etc 


65,000 




190 


70 


95 


40 






1055r 


8 686 


314 




186 5 


5 56 


071 


Similar to 30% nickel silver 














180 


180 










185 


5 61 


0071 


but of lower resistance 








73 


92 


33 




180 


1075t> 


8 691 


31 




165 6 


6 26 


081 


For silver-plated ware, spin- 
































ning, drawing, and for work 
































where a low percentage of 
































nickel is required 








7t 








180 


10306 


8 631 


31 










For white metal tubes, sheets, 
































wire, etc Also in plumbing 
































and decorations. 






106 




88 






180 
















For watch parts, etc Free 
































cutting 




34,000 






82 


32 






1010 


8 675 


31 




125 5 


8 27 


110 


For cheaper grades of silver- 


















960 B 








86 5 


11 99 


140 


plated ware. 






176 


6 


93 


20 






1060* 


8 176 


29 




58 6 


17 69 


U 198 


For diaphragms to withstand 
































pressure, also for its color 






18 


6 


99 


30 






I040< 


7 806 


28 


17 


70 08 


14 80 


173 


For diaphragms to withstand 
































pressure, also for resistance to 
































ordinary corrosion and wear 






190 


7 


100 


52 








7 746 


280 




95 1 


10 9 




For strength and resistance to 
































ordiiur> corrosion and wear 






19 


IOC 


100 


65 






(040t 


7 576 


27 




76 80 


13 5 


157 


For strength and resistance to 
































ordinary corrosion and wear 


45,000 




140 












10426 


7 585 


27 


16 


82 2 


12 61 


144 


For valve seats in airplane en- 
































gines and at elevated tem- 
































peratures 



h Jenkins and Hanson constitution diagram 
j Average linear coefhuent per cleg cent from 
25 to 300 deg cent Tests on rod Scien- 
tific Paper 410, U S Bureau of Standards. 
k At 18 1 deg cent 
m Cold worked and heat-trontod 
n Guertler-Tamrnann constitution diagram 
p Annealed, quenched, and heut-treated 



r Smith constitution diagram 

t Stockdale constitution diagram 

u G-cal per sec per sq cm per deg cent per 

cm at ?3 deg cent 
t T'ifel constitution diagram 
x Htvuor and Ilansen constitution diagram 
y Hard at 25 deg cent 
2 Heycock-Neville constitution diagram 



428 



THE USES OF COPPER 



TABLE 8. CHEMICAL AND PHYSICAL PROPERTIES 

(Variations must be 



Material 




Approximate Composition, 
per cent 


Tensile 
Strength, 
Ib per 8q in 


Elonga- 
tion in 
2 in, 
per cent 


Yield Point, (/>) 
Ib per t>q in. 


Johnson's 
Elastic Limit, 
Ib per sq m 


"<P~ 

*3 1 

wx 
















"S 9 
a o* 




E 


i.& 




j* 








^ 




o 




^ 




TJ 




1 * 
?jJ 







lu< 


.3 


% 








K 


% 


K 


3 


W 





& 


5 












Man- 


Al 


Iron 


Tin 




















Calsun bronze 


W 


95 50 




ga- 


2 50 




2 00 


135,000 


50,000 


4? 


35<- 






81,000 














nese 


























Manganese bronze 


R 


57 00 


40 00 


10 




1 45 


1 45 


90,000 


65,000 


15 


45 












Manganese bronze 


R 


59 00 


39 00 


50 




80 


70 


85,000 


60,000 


20 


45 












Everdur (A) 
1010 


S 
II 


96 00 
96 00 




1 00 
1 00 


Si 
3 00 
3 00 






113.000 
95,000 


55,000 
55,000 


5 
15 


48 
85 


75,000 
75,000 


20,000 
20,000 


72.000 
67,000 


9,300 


15 




W 


96 00 




\ 00 


3 00 






145,000 


59,000 




50. 


95,000 


25000 


90,000 






Everdur (B) 
1015 


T 
R 

S 


98 25 
98 25 
98 25 




25 
25 
25 


I 50 
1 50 
1 50 






65,000 
70,000 
70,000 


40,000 
40,000 
40,000 


15 
6 
6 


60 
60 
46 


60,000 
65,000 


10,000 
10,000 
10,000 


38,000 
















Cad- 
mium 
























Hitenso A 


\\ 


99 35 






65 






75,000 




3c 




47,000 




45,000 




15 6 




S 


99 35 






65 






54,000 




5 














Hitenso BB 


S 


99 00 






1 00 






60,000 


35,000 


3 


50 






36,000 








w 


99 00 






1 00 






92,000 


35,000 


3* 


50, 






55,000 






Hitenso C 


S 

\\ 


98 60 
98 60 






80 
80 




60 
60 


99,000 


36,000 
40,000 


4? 


50 

45e 




15,000 


59,000 












Iron 


Xi 


\1 


Mn 






















Tempaloy 917 


R 


81 90 


2 50 


5 00 


9 60 


1 00 




100,000 




10 




50,000 










Extrudid architec- 
tural bronze shapes 




57 00 


Zinc 
40 00 


Luul 
2 50 


Tin 
34 


Iron 
16 




70,000 


50,000 


10 


20 




















Be 


























Beryllium copper 


S 


97 40 




2 25 








118,000 


70,000 


4 3 


45 


105,000 


31,000 


79,000 


18,000 


17 2 










Ni 


































35 




























S 


97 40 




2 25 








192,000 


175,000 


2m 


6 3p 


138,000 


114,000 


130,000 


87,000 


18 4 










\*i 








m 


P 






m 


P 


m 


P 


m 










35 



























* For some alloys the figures given are for a 
temper slightly different from that commonly 
Kno^n as "hard " 

t Compared to water at 4 deg cent. 
J Soft 
R Rod 
S Sheet 
TTube 
W Wire 



a Temper not known 
6 Determination 

c Circular 7.'i, U S Bureau of Standards 
d Scientific Paper 410, U S Bureau of Stand- 
ards 

e Elongation of wire, per cent in 10 in. 
/ Corning GlaaH Works 

g Yield point taken as the load producing an 
extension under stress of 75 per cent 



COPPER-BASE ALLOYS 



429 



OF VARIOUS WROUGHT COPPER-BASE ALLOYS Continued 

expected in practice) 





Bnnell 


Rockwell 










a 


Electrical 








Hard- 


Hard- 












Properties 






Shearing 


nessNo 


ness No 


Bend 



W) 




2 


rt 


at 70 C 


^ 




Strength 


10-inm 


' B" 








3 


D. 








Ib per q in 


Ball, 
500-kg 


1/16-m 
Ball, 


dcg 


-a 


rt 


s, 


W 


5 


^ 


$ 


Uses Remarks 




Ixud 


100 U 




^ 







t^ -IT 1 


~5 


S-go. 







^ 




, 




9 




^ 




c 




-* 




5 i 


% 0^- 


ll 




"S 




"O 


^ 


7 




7 






o 


5j 


^ V 


S2.S 


~ 5"^ 


C o 




A 


& 


W 


& 




& 













Q 


U 





5"" 


H 




















10546 


8 5406 


308 




61 


17 




For electrical uses where cor- 
































rosion resistance is important 






170 


























I 1 or structural work due to 
































strength and resistance to 
































corrosion 










90 










8 3706 


302 




42 15 


24 6 


241 


1-or structural work due to 
































strength and resistance to 
































corrosion 






200 


70 


95 


40 




180 


10196 


8 5396 


308 


180 


155 


6 7 


078 


For strength and resistance to 


56,000 


33,000 


190 


60 
























corre)Sion Has strength of 


75,000 


35,000 




























mild stetl and corrosion re- 
































sistance of copper Welding 
































rod 










75 


20 




180 


I055r 


8 7406 


316 










I or str< ngth and resistance to 
















180 










86 4 


12 


129 


corrosion, bolt stock, and 










80 


3 




180 
















shttt metal requiring high 
































ductility 
















180 


1080/1 


8 896 


3212 




12 20 


85 




1 or electrical wire and cable, 






95 




62 






















etc 






98 




65 






180 


1076/i 


8 896 


3212 




12 95.v 


80 Oj/ 


824 


For electrical wire and cables, 


























I295j/ 


80 V 




contact fingers, commutator 
































segments, i tc. 
















180 


10706 


8 896 


3212 




18 85v 


55 O v 


556 


For electrical wire and cables. 


























18 85 V 


55 0^ 




etc 


















10546 


7 569 


273 










For strength and resistance 
































to corrosion 


















8846 


8 4326 


305 










For architectural shapes. 










102 


65 to 






9555 




297 


170 




I7p 




For springs, diaphragms, low 












73 










01 










dut> bushmgb and bearings, 
































and Bourdon tubes 










114 


112 5 










297 






18 to 


25p 


High resistance to fatigue. 






m 


P 


m 


P 










01 






25, n 


20m 





h Jenkins and Hanson constitution diagram 
; Average linear coefficient per deg cent from 
25 to ,500 des cent Tebts on rod Scien- 
tific Paper 110, U S Bureau of Standards 
k At 18 1 dog cent 
m Cold worked and heat-treated 
n Guertler-Tammann constitution diagram. 
p Annealed, quenched, and beat-treated. 



r Smith constitution diagram 

t Stockdale constitution diagram 

u G-cal per hec per b<j cm per deg cent per cm 

at 20 deg cent 

t Tafel constitution diagram 
x Bauer and Hansen constitution diagram 
y Hard at 25 deg cent 
2 Ueycock-Neville constitution diagram 



Table reprinted by permission from Handbook of Engineering Fundamentals, by O W Eshbach, pub- 
lished by John Wiley and Sons, Inc , New York, 19.JG 



CHAPTER XII 
PRODUCTION OF COPPER 

INTRODUCTION 

In this chapter we shall give a brief analysis of the production of 
copper throughout the world. Copper and copper alloys were known 
to man long before there was any written history, but until 200 to 300 
years ago the supply of copper (and other metalb) available for man's 
use was insignificant when compared with the amounts in use today. 
Kings, nobles, and other wealthy people possessed most of the metal, 
and copper was classed with the precious metals in value Mining 
was confined to rich ores found near the Mirface, and metal production 
was limited. As far a> any considerable production of copper is 
concerned, history does not begin until about 1800 

With respect to the world production of copper in 1800, Julihn 1 makes 
the following statement: 

In 1800 theie \\a^ a^ \et no established production from North America, 
Africa or Australasia, but Europe produced an average of about 12,400 tons 
a year, including about 7,300 tons fiom Cheat Britain, 3,300 ton^ from Rus- 
sia, and 1,700 tons from S\\eden, Xonvav, and (Jermanv Japan produced 
about 3,100 ton*, a vear and South Amenra about 2,600 tons a rear 1,700 
tons from Chile and 900 tons from Venezuela All other production appears 
to have been casual in character and dight in quantity 

Comparing the world production in 1800 (18,100 tons) with the 1937 
production (2,343,156 metric tons) - we find that the yearly production 
has increased over 130-fold in a period of 137 years Some of the 
oldest known deposit^ are still producing; in 1938 Spam produced 
about 30,000 metric tons of copper, and the island of Cyprus 29,780 
tons. Copper has been mined in both of these places since the earliest 
times , in fact our word " copper " is derived through Greek and Roman 
sources from the name Cyprus. It is interesting to note that although 
the present production from Spain and Cyprus is each only a little 

1 Julihn, C E, Summarized Data of Copper Production- U. S. Bur. Mines 
Econ Paper 1, p 30, 1928 

2 Minerals Yearbook, 1939, p. 115, U. S Bur. Mines. 

430 



COMPARISON WITH OTHER METALS 431 

over 1 per cent of the world production, yet either of these is now 
producing more copper than the entire world production in 1800. 

The present large production of copper is the result of the great 
demand for the metal which followed the industrial revolution in the 
nineteenth century. Two primary factors are responsible for the large 
increase in production the discovery of immense new ore deposits 
in Africa and the Americas, and the development of techniques for 
winning copper from low-grade ores 

COMPARISON WITH OTHER METALS 

Copper ranks second in tonnage of all metals produced; iron exceeds 
it both in tonnage and in value, and in recent years the value of gold 
produced has exceeded that of copper. Statistics of world production 
of the nine most important metals are given in Table 1. 

Although the amount of metal produced in the world depends on 
many economic, political, and military factors, so that the production 
fluctuates from year to year, there are certain generalizations which 
apply to the position of copper m relation to other metals, most of the 
points listed below are illustrated by the statistics m Table 1. 

1. Copper is the most important non-ferrous base metal, both with 
respect to the tonnage produced and to its value. It has maintained 
this position for many years and probably will maintain it for many 
years to come. 

2. In recent yeais the higher price of gold has stimulated production, 
and now the value of gold produced greatly exceeds that of copper. In 
1929, however, the value of the copper produced was almost twice that 
of gold. 

3 From the tonnage standpoint, there is from 40 to 50 times as 
much pig iron produced as copper Copper, however, is worth from 
10 to 20 times as much as pig iion, so that the ratio of the total value 
of these metals is much smaller than the tonnage ratio 

4. Lead and zinc are produced in tonnages only slightly less than 
that of copper (Table 1), but the prices of both these metals are con- 
sistently much lower than the copper price. 

5. In recent years the production of both aluminum and nickel has 
increased rapidly; aluminum rose from eighth place in 1929 to fourth 
m 1938 (Table 1). In the same period the tonnage of nickel produced 
was doubled Aluminum tonnage is still well below the tonnage of 
copper, lead, or zinc, but in value the aluminum ranks next to copper 
(1938). 



432 



PRODUCTION OF COPPER 



TABLE l a 
WORLD PRODUCTION OF METALS FOR 1929, 1936, AND 1938 

1929 



Metal, Ranked 
According to 
Value 


Production 
(short 
tons) 


Pi ice 


Value 


1. Pig iron 


107,408,000 


$20 00/long ton 


$1,918,000,000 


2. Copper 


2,104,110 


18 107 cents /pound 


757,479,600 


3. Gold 6 


670 


$20.67/troy ounce 


403,366,000 


4 Lead 


1,935,110 


6 833 cents/pound 


264,300,000 


5. Zinc 


1,621,230 


6.512 cents/pound 


211,100,000 


6. Tin 


213,143 


45 19 cents/pound 


192,700,000 


7. Silver' 


9,000 


53.3 cerits/troy ounce 


147,212,439 


8. Aluminum 


298,000 


24 cents /pound 


143,000,000 


9 Nickel 


62,300 


35 cents /pound 


43,610,000 



1936 



1 Pig iron 


100,578,210 


$18 90 /long ton 


1,700,000,000 


2. Gold 6 


1,219 


$35 00/troy ounce 


1,233,312,825 


3. Copper 


1,836,480 


9 474 cents /pound 


350,000,000 


4 Tin 


199,034 


46 42 cents/pound 


185,000,000 


5. Aluminum 


394,700 


20 5 cents/pound 


161,800,000 


6. Zinc 


1,646,786 


4 901 cents/pound 


161,400,000 


7. Lead 


1,642,726 


4 710 cents/pound 


158,500,000 


8. Silver" 


8,557 


45 399 cents/troy ounce 


114,007,000 


9. Nickel 


98,400 


35 cents/pound 


68,880,000 


1938 


1. Pig iron 


90,000,000 


$21 67/long ton 


1,745,000,000 


2. Gold 6 


1,263 


$35 00/troy ounce 


1,289,785,000 


3. Copper 


2,185,000 


10 00 cents/pound 


437,000,000 


4. Aluminum 


648,000 


20 cents/pound 


259,200,000 


5. Lead 


2,070,000 


4 74 cents/pound 


196,300,000 


6. Zinc 


1,750,000 


4 610 cents/pound 


161,200,000 


7. Tin 


165,000 


42 26 cents/pound 


124,500,000 


8. Silver 


9,050 


42 944 cents/ troy ounce 


113,496,269 


9. Nickel 


123,500 


35 cents/pound 


86,450,000 



a Mineral Industry, 1929, 1936, and 1938, McGraw-Hill Book Co , New York The world price of 
silver is used, in other cases the price is the average United States price for the year 

6 Production of gold in troy ounces 19,500,000 in 1020, 35,500,000 in 1936, and 36,851,000 in 1938 
c Production of silver in troy ounces 262,000,000 in 1929, 250,000,000 in 1936, and 264,289,000 in 
1938 



WORLD PRODUCTION OF COPPER 



433 



WORLD PRODUCTION OF COPPER 

The data in Table 2 show the total world production of new copper 
from 1800 to 1940, taken by decades, and these data are plotted in 
Figure 1. Considering the production data in this way gives a better 
picture of the general trend than a direct plot of annual production 
(Fig. 2) because the irregularities are smoothed out. A curve drawn 
through the extremities of the ordmates in Figure 1 shows the general 
characteristics of world production. The total production in the 
decade 1931-1940 is almost 100 times the 1801- -1810 production, and 
the most rapid increase in production took place from 1881 to 1920. 
For the hundred years 1821 to 1920 the average increase was 52.8 per 
cent per decade. Since 1920 the curve has been lising less steeply but 
still does not appear to have " flattened out." 

TABLE 2 a 
WORLD PRODUCTION OF COPPER BY DECADES 



Decade 


Production 
(short tons) 


Increase 
(per cent) 


Decade 


Production 
(short tons) 


Increase 
(per cent) 


1801-1810 


182,000 




1871-1880 


1,423,744 


23 9 


1811-1820 


188,496 


3 6 


1881-1890 


2,488,591 


74.8 


1821-1830 


273,504 


45 1 


1891-1900 


4,149,353 


66.7 


1831-1840 


364,448 


33 3 


1901-1910 


7,628,334 


83 8 


1841-1850 


493,808 


35 5 


1911-1920 


12,187,341 


59 8 


1851-1860 


759,079 


53 7 


1921-1930 6 


15,198,926 


24 9 


1861-1870 


1,149,344 


51 4 


1931-1940 C 


17,000,000 


11 8 



U S Bur Mines Econ Paper No 1, 1928 

b Mineral Industry for 1038, McGraw-Hill Book Co , New York 

c Estimated 

Figure 2 shows the yearly world production in the period 1881-1938. 
Since 1910 the production curve has shown a series of violent fluctua- 
tions, although the general trend is the same as that noticeable in 
Figure 1. In 1937 the world production reached an all-time high of 
about 2]/i million short tons; previous records were 1.5 million tons in 
1919 and 2.1 million tons in 1929. 

It is very difficult to predict what the future world production of 
copper will be because there are many unknown factors which will 
affect it. It is possible to extrapolate the production curve into the 
future by drawing " average " curves or '* trend lines," and these, of 
course, indicate that copper production will increase rapidly according 
to the sharp rise of the production curve drring the last 50 years 
(Figs. 1 and 2). Such predictions indicate an annual production of 



434 



PRODUCTION OF COPPER 



7 to 8 million tons in the early part of the twenty-first century three 
times the 1937 production. An increased demand such as this would 
require the discovery of new ore deposits much greater than the reserves 




I 1801 11811 I 1821 I 1831 I 1841 I 1851 I 1861 I 1 8 7 1 M 88 iTl 89 TT 1 901(191 IfT 92 1 | 1 93 1 
I 1810 I 1820 | 1830 I 1840 | 1850 I 1860 | 1870 | 1880 I 1890 I 1900 I 1910 | 1920 | 1930 I 1940 



(Data from Table 2) 

FIG. 1 World Production of Copper by Decades 

we have today, and also the mining, milling, and smelting methods 
would have to be completely revolutionized in order to exploit low- 
grade and complex ores. 

There are many other factors which must be considered which cast 
some doubt on the accuracy of a simple extrapolation of the trend line 
to determine future production. 

1. An increasing demand in the next century corresponding to the 
increase in the past 50 or 60 years will follow only if there is a 
corresponding increase in industrialization throughout the world. If 
China, India, and other countries were to develop an industrial plant 
comparable to that of the United States, Great Britain, or Germany, 
then the copper production would probably exceed the predictions of 
the most optimistic statistician. Some such development seems likely, 
but its nature and extent cannot be predicted. 

2. The figures we have quoted thus far refer to primary copper or 
copper obtained from newly mined ores. Much of the world's con- 
sumption, however, is secondary or scrap copper. Up to the end of 



WORLD PRODUCTION OF COPPER 



435 



1938 the total world production of primary copper was about 60 million 
short tons and at present the world is adding to this stock at the 
rate of about 2 million tons yearly. Copper is an " indestructible " 
metal, and much of the copper used commercially can be reclaimed 
and used over and over again. This rapidly growing supply of copper 
is found to decrease the demand for newly mined copper, and the 
time may come when the production of primary copper will be just 
sufficient to replace the unavoidable loss or wastage in the circulating 




1890 



1900 



1910 



1920 



1930 



(Data from Econ Paper No. 1, U. S Bur Mines, and The Mineral Industry During 1938, 

McGraw-Hill Book Co , New York) 

FIG. 2. World Annual Production of Copper. 

supply of metal. Secondary copper is already an important factor in 
world markets, and we shall have more to say on this subject in 
another section. 

3. For certain uses other materials are being substituted for copper 
aluminum, nickel, and some non-metallic substances. This may 
eventually prove to be an important factor; but, as we have noted, it 
is unlikely that a substitute can be found for copper where the 
material must have high electrical and thermal conductivity. 



436 



PRODUCTION OF COPPER 



8 

z 
& 

S 

O 

E 

s 

CO 

& 

O 
E 

ti 

g 



W 2 

W g 

^ g 

S 6 

* 

n 



s 

0< 

<S 

fe 
O 

o 



&5 







O5 (N CO GO rH 



O O C<J Tt< CD 

i i CO 00 TH C5 
O5 T^ (N TH CO 



T^ (M rH l> <> O 

SO rH UQ <N rH 
1 14 T 1 



O CO M f-( ( 



O 
(M 



^H O O> Tfl 

t^ GO O CO 

O" O~ OC~ rH~ of 
C5 GO CO CO t>- 



00 CO 

10" GO" o 

l^ CO O5 

co" IN" co" 



S 08 
*g *C 

4l 



rH CO ^ 

CD Ca O 

LQ I^O Tf 

" "8" 



oj 
cj 



8 



8 



E2 



W 



WORLD PRODUCTION OF COPPER 



437 



4. Although the future may create greater demands for copper, it 
is possible that present reserves may become exhausted and that new 
discoveries of copper ore may not keep pace with the demand. In such 
a case the price of copper would rise, and pressure of this sort would 
result in a greater use of copper scrap and more intensive development 
of substitutes for copper. 

Table 3 lists the copper production of the continents for various 
periods from 1800 to 1938, these data are plotted in Figure 3. From 





- 


22 


pn Australasia 


- 


20 


HI Africa 
S Asia 


- 


1.8 


iH Europe ^B 

South America ^H 


- 


1.6 1 


[_J North America H^^ 

Hi 


- 


145 

c 
o 


Horizontal distance -Time 

Vertical distance Average 
yearly production 

Area---Total production 




k 




<\ 


- 


~ f- 

cn bo o "ro 

Millions of Short 1 




SISSIffi 






l]lllllllllllllllli 














- 


0.4 


Asia, Africa, & Australasia^ 










- 


0.2 


1801-1900 


1901-1925 


^-12-> -1*- 
1926 1937 1938 


(Data from Table 3) 



FIG. 3. World Production of Copper Since 1800. 

these figures we can compare the relative amounts of metal produced 
throughout the world and also the relative importance of the continents 
as producers of copper. 

The total world production of copper in the half century from 1801 
to 1850 is given in Table 4. The figures in this table show several facts 
which may be contrasted with present day figures 

1 England was the most important copper producing country; 
today the amount of primary copper produced in England is negligible. 

2. Europe was the largest producer of all the continents. 

3. No production whatever was shown for \frica, and the bulk of 
the North American production came from Cuba. All the production 



438 



PRODUCTION OF COPPER 




WORLD PRODUCTION OF COPPER 



439 



WORLD COPPER PRODUCTION, 1801-1850 


8, 
e 

H 


|| 


iQ Oi O CN 

oT ad* 8 CD 


1 


I 

OJ 


NT CO^ CO 

OS O -i 


VI 

I 

t-t 
O 

2 
^o 




<N O 
g 




> 

a 

i 

i 

02 




f 

HH 


tC 00 




Production 
Yearly average produc- 
tion 




Production, short tons 
Yearly average, short tons 
Per cent of Asiatic total 
Per cent of world total 


1| 


t^ ^ 

00 00 


Austria- 
Hungary 


f o 
51 co oo i-( 


North America 


Total 
North 
America 


"M O O CD 


I 


b- i-H . JN 

00_ "7 '- *- 


United 
States 


! * :: 


South America 


Total 
South America 


03 Tf 


Germany 


00 CD 

382 


Xj 

6 


1 !- "! 


Sweden 


iO OO 1 s "* 
O O ^ ^ 

o 1 10 co 


Venezuela 


|C oc co 




Production, short tons 
Yearly average, short 
tons 
Per cent of North 
American total 
Per cent of world total 


'53 
55 


CN ^ ^ 


U 


S ^ 


England 


CO CO 

cb T-H ^ co 


Australasia 


O to | 


00 iO l> 

>o co i-< 




Production, short tons 
Yearly average, short tons 
Per cent of South American total 
Per cent of world total 




Production, short tons 
Yearly average, short tons 
Per cent of European total 
Per cent of world total 




Production, short tons 
Yearly average, short 
tons 
Per cent of world total 



440 PRODUCTION OF COPPER 

credited to the United States came in the decade 1841-1850, but Cuba 
had been producing copper since before 1830. 

4. South America accounted for a good share of the world's produc- 
tion, and the bulk of the South American copper came from Chile since 
1800. Chile has always produced a considerable share of the world's 
copper (Fig. 2). 

5. The average annual production was small when compared to 
today's figures, and the total world production in the 50 years from 1801 
to 1850 w r as only about 60 per cent of the world production in the single 
year of 1937. 

EUROPE 

Until 1800, Europe was far and away the leading producer of copper 
in the world, and in the first half of the nineteenth century (1801-1850) 
it accounted for 63 per cent of the world's copper. Since that time 
the relative importance of European copper has steadily declined; the 
yearly tonnage has increased, but Europe's share of the world total 
has dropped to 8 or 10 per cent because of the great increase in pro- 
duction in other parts of the world. 

Spain and Portugal. The main producing districts in Spain are Rio 
Tinto and Tharsis, and most of the Portuguese production comes from 
the Mason and Barry mines. From 1801 to 1927 inclusive, Spain and 
Portugal constituted the leading European producer of copper and in 
that period accounted for 697 per cent of the world's production. 3 
Table 5 gives the copper production of Spain and Portugal for various 
periods; note that in recent years its importance has declined both with 
respect to European and world production. There has been relatively 
little fluctuation in the yearly tonnage produced. 

Germany. The copper production of Germany is shown in Table 6 
for the same periods as those shown in Table 5. During the periods 
shown, Germany has generally accounted for one-fifth to one-fourth 
of the European production, and the yearly tonnage has shown a 
steady increase up to 1937. 

The most important deposits in Germany are the low-grade copper- 
bearing shales of Mansfeld in central Germany; these account for 
practically all the present German production. 

Russia. Table 7 gives the copper production of Russia during 
periods comparable with those in Tables 5 and 6. Note particularly 
the rapid increase in tonnage in the last few years; in 1937 Russia 
accounted for 38.4 per cent of the European production and about 4 per 

3 U. S. Bur Mines Econ. Paper No. 1, 1928. 



EUROPE 



441 



TABLE 5 a 
COPPER PRODUCTION OF SPAIN AND PORTUGAL 



Year or 
Period 


Production 
(short tons) 


Average 
Yearly 
Production 


Per Cent 
of European 
Total 


Per Cent 
of World 
Total 






(short tons) 






1871-1880 


219,408 


21,941 


49.9 


15.41 


1881-1885 


241,142 


48,228 


60 6 


21.81 


1886-1890 


299,870 


59,974 


66 


21.69 


1901-1910 


563,267 


56,327 


55.2 


7.39 


1911-1920 


448,849 


44,885 


38 2 


3.68 


1921-1925 


205,916 


41,183 


50 7 


3.49 


1927 


60,351 


60,351 


43 3 


3 60 


1929 6 


53,500 


53,500 


31 4 


2 53 


1933 6 


40,500 


40,500 


20.6 


3.54 


1937 6 


34,500 


34,500 


13 


1.38 



U S Bur Mines Econ Paper No 1. 
6 Mineral Industry in 1938, Vol 47 



TABLE 6 
COPPER PRODUCTION OF GERMANY 



Year or 
Period 


Production 
(short tons) 


Average 
Yearly 
Production 


Per Cent 
of European 
Total 


Per Cent 
of World 
Total 






(short tons) 






1871-1880 


73,360 


7,336 


16.4 


5 15 


1881-1885 


82,951 


16,590 


20 8 


7.50 


1886-1890 


89,098 


17,820 


19.6 


6.44 


1901-1910 


241,791 


24,179 


22 4 


3 18 


1911-1920 


375,641 


37,564 


32 


3 08 


1921-1925 


107,648 


21,530 


26.5 


1 82 


1927 


31,306 


31,306 


23.3 


1.87 


1929 6 


32,100 


32,100 


18.8 


1 51 


1933 6 


34,200 


34,200 


17.4 


2 90 


1937 6 


36,200 


36,200 


13.6 


1 44 



a U S Bur Mines Econ Paper No 1 
6 Mineral Industry in 1938, Vol 47 

cent of the world total. In these tabulations the entire Russian pro- 
duction is considered as European production. 

The following quotation on the ore deposits and reserves of the 
U.S.S.R. are taken from The Mineral Industry. 4 These in turn were 

4 The Mineral Industry During 1938: Vol. 47, p. lol, McGraw-Hill Book Co, 
New York. 



442 



PRODUCTION OF COPPER 



TABLE 7 a 
COPPER PRODUCTION OF RUSSIA 



Year or 
Period 


Production 
(short tons) 


Average 
Yearly 
Production 


Per Cent 
of European 
Total 


Per Cent 
of World 
Total 






(short tons) 






1871-1880 


36,960 


3,696 


8.3 


2 60 


1881-1885 


22,677 


4,535 


5.7 


2 05 


1886-1890 


26,258 


5,252 


5.8 


1 90 


1901-1910 


142,081 


14,208 


13.1 


1 86 


1911-1920 


215,406 


21,541 


18.4 


1.77 


1921-1925 


19,146 


3,829 


4.6 


0.32 


1927 


13,227 


13,227 


9.5 


0.79 


1929 6 


28,400 


28,400 


16.6 


1 34 


1933 6 


36,000 


36,000 


18 4 


3 15 


1937 6 


102,000 


102,000 


38 4 


4 07 



a U S Bur Mines Econ Paper No 1 
6 Mineral Industry in 1938, Vol 47. 

taken from the London Mining Journal, Tsvetme Mctalli, No 11, 1937, 
and Der Out-Express (Berlin), December 1937. 

The most extensive ore deposits occur in Kazakstan, the tuo main groups 
being the Kounrad ores near Karsakpai and the Balkash ores near the 
northern shore of Lake Balkash. The Urals also certain considerable de- 
posits of copper ore, and other deposits occur in Uzbekistan, Bashkiria, Mid- 
dle Volga, West Siberia, Transcaucasia, Leningrad Province, and the Kola 
Peninsula. 

Construction of new works seems to be proceeding very slowly and some 
of the " Giants " projected some years ago still exist only on paper Oi 
six important new works, under construction or projected, the Balkash 
plant, designed to produce 100,000 tons a year, and on which 300 million 
roubles have been spent since 1930, has not yet produced copper The 
Jezkazgan plant at Karsakpai in Kazakstan, \\hich is intended to work 
on low-grade ores and to produce eventually 200,000 tons of black copper 
per annum, will be three or four years before it \\ill begin to produce 

Various decrees, regulations and conferences have attempted to deal with 
the problem of inefficiency in the last few years and a conference at Sverd- 
lovsk found that the Ural works, which produce about four-fifths of the 
present copper output, had only fulfilled about 50 to 60 per cent of their 
programme in the first nine months. 

U.SS.R reserves of metallic copper in ore were estimated on January 1, 
1935, to be 10,635,000 tons, distributed as follows: * * * 



ASIA 443 

Tons 

Kazakstan 6,404,400 

Urals 2,116,500 

Uzbekistan 1,285,400 

Bashkiria 330,500 

Middle Volga 324,500 

West Siberia 108,600 

Caucasus 52,400 

Leningrad Province 8,300 

Karelia 4,900 

Another quotation 5 from Bulletin 36 of the Imperial Institute 
(London), Vol 1, January-March 1938, is as follows: 

It will thus be seen that, adding present production of copper to the 
scheduled output of works under construction or proposed to be constructed 
in the next few years, the Soviet Union is planning for an eventual output, 
say within the next ten >ears, of about 500,000 tons of copper per annum. 

Yugoslavia. The principal production of copper in Yugoslavia 
conies from the Mines de Bor, which, operated under French control, 
reported the production of 41,992 metric tons of copper in 1938 com- 
pared with 39,410 in 1937. An electrolytic refinery was completed 
and opened on July 2, 1938, and by the end of the year the production 
of this electrolytic plant was reported to be 1000 tons per month. 6 

Yugoslavia did not produce much copper previous to 1920, but 
since then it has become one of the important European producers 
(Table 8) 

Norway, Sweden, Finland. The combined production of Norway, 
Sweden, and Finland i<* shown in Table 8. Most of this production is 
from Norway, although the Finnish production has increased in 
recent years 

Other European Countries. Small amounts of copper are produced 
in Austria, France, and the Island of Cyprus. Great Britain, which 
formerly (Table 4) was the world's leading producer (from mines in 
Cornwall and Devon), no longer produces an appreciable amount of 
copper by 1926 England produced only 0.01 per cent of the 
world's total. 

ASIA 

Japan is the principal copper producer in Asia, as may be seen 
from Table 9. Practically all the recorded Asiatic production previous 

6 Minerals Yearbook, 1939, p 123, U. S. Bur. Mines. 
6 Minerals Yearbook, 1939, p. 123, U S Bur. Mines. 



444 



PRODUCTION OF COPPER 



TABLE 8 a 
COPPER PRODUCTION OF YUGOSLAVIA 



Year or 
Period 


Production 
(short tons) 


Average 
Yearly 
Production 


Per Cent 
of European 
Total 


Per Cent 
of World 
Total 






(short tons) 






1911-1920 


4,476 


448 


0.31 


0.03 


1921-1925 


34,930 


6,986 


8.6 


0.59 


1927 


14,220 


14,220 


10.2 


85 


1929 & 


22,800 


22,800 


13.3 


1.07 


1933* 


44,100 


44,100 


24 8 


4.25 


1937 b 


43,400 


43,400 


18 


1 91 



COMBINED COPPER PRODUCTION OF NORWAY, SWEDEN, AND FINLAND 



1871-1880 


32,032 


3,203 


7.2 


2 25 


1881-1885 


19,139 


3,828 


4 8 


1.73 


1886-1890 


13,560 


2,712 


3 


98 


1901-1910 


63,680 


6,368 


5 9 


83 


1911-1920 


62,432 


6,243 


5.4 


0.52 


1921-1925 


6,715 


1,343 


1.6 


0.11 


1927 


13,227 


13,227 


9.5 


0.79 


1929 6 


24,700 


24,700 


14 5 


1 07 


1933 6 


32,400 


32,400 


16 5 


2 83 


1937 6 


42,200 


42,200 


17 5 


1 85 



U S Bur Mines Kcon Paper No 1 
6 Mineral Industry in 1938, \ol 47 

to 1900 was credited to Japan, and in the period 1900-1925, Japan 
produced more than 98 per cent of all Asiatic copper. In recent years 
the proportion produced by Japan has decreased (Table 9) owing to 
increased production by India, Turkey, and China. 

Japan. The main producing districts of Japan are Besshi, Furu- 
kawa, and Ashio, all of which produce about the same grade of 
copper. 7 Statistics for Japanese production are given in Table 9. 

AUSTRALASIA 

Australasia has been a steady producer of copper for many years 
(Table 10) ; it has never been a large producer, however, and in 
recent years its production has been less than 1 per cent of the 
world's total. 

Australasian production includes that from the islands of New 
Zealand, Tasmania, and Papua as well as from the Australian main- 

7 U S Bur. Mines, Econ. Paper No 1. 



AUSTRALASIA 



445 



land; the important districts are Mount Lyell in Tasmania and Mount 
Morgan, Mugana, Chillogee, and Wallaroo in Australia. In recent 

TABLE 9 a 
COPPER PRODUCTION OF JAPAN 



Year or 


Production 


Average 
Yearly 


Per Cent 
of Asiatic 


Per Cent 
of World 


Period 


(short tons) 


Production 


Total 


Total 






(short tons) 






1871-1880 


38,080 


3,808 


100 


2 68 


1881-1885 


41,068 


8,214 


100 


3.71 


1880-1890 


75,964 


15,193 


100.0 


5.49 


1901-1910 


430,883 


43,088 


97.5 


5 65 


1911-1920 


854,326 


85,433 


99.5 


7.01 


1921-1925 


329,664 


65,933 


96.5 


5.58 


1927 


69,872 


69,872 


91.6 


4.17 


1929* 


83,300 


83,300 


90.5 


3 92 


1933 6 


76,100 


76,100 


86 5 


6.65 


1937 6 


83,500 


83,500 


63.9 


3.34 



U S. Bur Mines Econ Paper No 1. 
6 Mineral Industry in 1938, Vol 47 



TABLE 10 
COPPER PRODUCTION OF AUSTRALASIA 







Average 


Per Cent 


Year or 


Production 


Yearly 


of World 


Period 


(short tons) 


Production 


Total 






(short tons) 




1871-1880 


123,648 


12,365 


8 68 


1881-1885 


63,037 


12,607 


5.70 


1886-1890 


51,506 


10,301 


3.73 


1901-1910 


377,219 


37,722 


4.94 


1911-1920 


427,932 


42,793 


3.51 


1921-1925 


75,685 


15,137 


1 28 


1927 


12,800 


12,800 


77 


1929 6 


16,000 


16,000 


0.76 


1933 6 


17,600 


17,600 


1.54 


1937 6 


22,000 


22,000 


0.88 



U. S. Bur. Mines Econ Paper No 1. 
6 Mineral Industry in 1938, Vol. 47. 



years the principal producer has been the Mount Lyell Mining and 
Railway Company, Ltd., Mount Lyell, Tasmania. 



446 



PRODUCTION OF COPPER 



AFRICA 

Africa today is one of the world's large producers of copper, but 
(Table 11) it is only in recent years that it has assumed an outstanding 
position. African production comes from three principal sources 
(1) Katanga in the Belgian Congo, (2) Northern Rhodesia, and (3) 
South Africa. The bulk of the production previous to 1910 came 
from South Africa, with a little from Algeria, but today it is Katanga 
and Rhodesia that account for most of the African copper. These 
two areas are adjacent, and together they constitute the most important 
copper producing province in the world. 

TABLE ll a 
COPPER PRODUCTION OF AFRICA 







Average 


Per Cent 


Year or 


Production 


Yearly 


of World 


Period 


(short tons) 


Production 


Total 






(short tons) 




1871-1880 


42,112 


4,211 


2 96 


1881-1885 


31,268 


6,254 


2.83 


1886-1890 


71,033 


14,207 


2.85 


1901-1910 


65,659 


6,566 


86 


1911-1920 


288,606 


28,861 


2.37 


1921-1925 


425,355 


85,071 


7 20 


1927 


120,763 


120,763 


7 22 


1929 6 


173,000 


173,000 


8 15 


1933 6 


200,500 


200,500 


17 5 


1937 6 


416,000 


416,000 


16 6 



U S Bur Mines Econ Paper No 1. 
6 Mineral Industry in 1938, Vol 47. 

In South Africa the most important copper producer is the Messina 
Development Company in the Transvaal. There is also some produc- 
tion from Southern Rhodesia, and in the early days there was some 
copper produced from Algeria. In recent years (Tables 12 and 13) 
better than 90 per cent of Africa's copper has come from Katanga and 
Northern Rhodesia. 

Katanga. The copper mines of the province of Katanga in the 
Belgian Congo are operated by the Union Miniere du Haut Katanga 
which was organized in 1910. The principal production has been 
from oxidized copper ores. We have already discussed the various 
metallurgical treatments that have been employed in exploiting these 
deposits. In recent years considerable sulfide ore has been mined and 
treated; the importance of the sulfide ores in this district will likely 



NORTHERN RHODESIA 447 

increase in the future. The brief history of the development of the 
Belgian Congo copper belt (up to 1936) which follows is taken from 
the Mineral Industry during 1936 ; 8 part of this material is quoted 
from the South African Mining and Engineering Journal. 

The Belgian Congo copper belt is some 10,000 square miles in area, 
being about 200 miles long and averaging 50 miles m width. Within this 
area no fewer than 200 separate potential copper mines exist, and of these 
less than 10 per cent have been worked Up to the end of 1935 about 
1,500,000 tons of metal had been produced from 25,000,000 tons of ore, and 
shareholders had received about 7,000,000 m dividends The principal 
concentrator plant is at Panda m approximately the middle of the belt, and 
the smelter at Lubumbashi, near Elisabethville. An electrolytic leaching 
plant has recently come into operation at Panda to treat the lower-grade 
ore. It is estimated that the developed ore reserves amount to 78,000,000 
tons, containing over 5,000,000 tons of copper. 

It was not until 1910, when the Union Mmiere had been formed and the 
Rhodesia Katanga Railway had reached Elisabethville, that vigorous develop- 
ment became possible The first two mines to be opened up were the 
Etoile, near Ehsabethville, which stands in the southeastern portion of the 
Katanga copper belt, and the Kambove, about 80 miles from the capital 
These were the only producers until the Ruashi was opened in 1922, fol- 
lowed by the Luishia and Likasi mines, which produced rich oxide ores from 
shallow deposits In 1926 the Kipushi mine, now known as the Prince 
Leopold, began production on a large scale from its very rich copper-silver 
sulphide ore. This mine is at present [1936] the largest producer. 

The Union Mmiere du Haut Katanga produced in 1936, under the inter- 
national curtailment agreement, about 130,000 metric tons This compares 
with the highest output of 139,000 tons m 1930. A new and elaborate devel- 
opment program has been elaborated which includes the opening up of the 
Sesa mine near Kamboroe and restarting of the reverberatory furnaces at 
Panda. The erection of a new roasting plant for its sulphurous ore is 
planned to be ready for operation towards the middle of 1938. 

The copper production of the Belgian Congo is given in Table 12 
Northern Rhodesia. Table 13 gives the production of copper in 
Rhodesia for a number of years. Previous to about 1930 this copper 
came from Southern Rhodesia, and it was not until about 1931 that 
the mines in Northern Rhodesia came into production. Since that 
time the copper fields of Northern Rhodesia have shown a rapid in- 
crease in production and are now producing considerably more than 
the rich deposits of Katanga (Tables 12 and 13) ; the Northern Rhodesia 
copper field is the greatest in the world. The following brief account 

8 The Mineral Industry during 1936, Vol. 45, pp 148-149, McGiaw-Hill Book 
Co., New York. 



448 



PRODUCTION OF COPPER 



TABLE 12 a 
COPPER PRODUCTION OF BELGIAN CONGO 



Year or 
Period 


Production 
(short tons) 


Average 
Yearly 
Production 


Per Cent 
of African 
Total 


Per Cent 
of World 
Total 






(short tons) 






1911-1915 


39,947 


7,999 


36 7 


0.74 


1916-1920 


124,141 


24,828 


69 


1.82 


1921-1925 


338,835 


67,767 


79 7 


5 73 


1926 


88,889 


88,889 


81 


5 46 


1928 6 


124,000 


124,000 


88 


6 55 


1930 6 


153,000 


153,000 


84.3 


8 82 


1932 6 


59,500 


59,500 


38 4 


5.97 


1934 6 


121,000 


121,000 


42 1 


8 70 


1936 6 


105,500 


105,500 


38 7 


5.71 


1937 6 


166,000 


166,000 


39 9 


6.62 



U S Bur Mines Econ Paper No 1 

6 The Mineral Industry m 1938, Vol 47 (Production of Union Minidre du Haut Katanga ) 

of the ore deposits of Northern Rhodesia is abstracted from an 
article by Bateman. 9 

The Northern Rhodesia copper belt lies adjacent to the boundary 
of the province of Katanga, Belgian Congo; the belt is about 140 miles 
long by 40 miles wide and trends in a northwesterly direction. N'Dola, 
on the main line of the Congo-Rhodesian Railway, is the distributing 
center from which branch lines extend to Roan Antelope (22 miles), 
Nkana (45 miles), and Mufulira (59 miles). 

The presence of oxidized ores in Northern Rhodesia had long been 
known, but these low-grade oxidized ores compared unfavorably with 
the rich ore of Katanga. Work on oxidized ore commenced at the 
old Bwana M'Kubwa mine in 1903; copper shipments started in 1913, 
but after intermittent operations this mine was closed. Sulfides were 
disclosed by boring at N'Changa but their significance was not ap- 
preciated because they were mixed with oxides; the true importance 
of the sulfides was not realized until the sulfide zone had been 
penetrated at Roan Antelope in November, 1925. Later the Nkana, 
Mufulira, Chambishi, Baluba, and Extension deposits were discovered. 
The Roan, Nkana, Mufulira, and Chambishi deposits consist almost 
entirely of sulfide ore. The N'Changa contains mixed sulfides and 
oxides, and the Extension is largely oxidized ore. 

9 Bateman, A. M., The Northern Rhodesia Copper Belt: in Copper Resources 
of the World, Vol. 2, p. 713; published by XVI Internat. Geol. Cong., Wash- 
ington, 1935. 



NORTHERN RHODESIA 



449 



TABLE 13 a 
COPPER PRODUCTION OF RHODESIA 



Year or 
Period 


Production 
(short tons) 


Average 
Yearly 
Production 


Per Cent 
of African 
Total 


Per Cent 
of World 
Total 






(short tons) 






1911-1915 


4,528 


905 


4.0 


08 


1916-1920 


17,155 


3,431 


9.5 


0.25 


1921-1925 


14,916 


2,983 


3 5 


25 


1926 


850 


850 


7 


05 


1931 6 


3,460 


3,460 


2 04 


23 


1932 6 


62,500 


62,500 


40 3 


6 26 


1933 C 


117,000 


117,000 


58 3 


10 2 


1935 d 


161,000 


161,000 


54 4 


9 95 


1936 d 


152,300 


152,300 


58 5 


8 63 


1937 6 


233,000 


233,000 


55.2 


9 34 



a U S. Bur Mines Econ Paper No 1 

6 The Mineral Industry during 1934, Vol 43 (Northern Rhodesia). 

c The Mineral Industry during 1933, Vol 42 (Northern Rhodesia) 

d The Mineral Industry during 1937, Vol 46 (Northern Rhodesia) 

The Mineral Industry during 1938, Vol 47 (Northern Rhodesia). 

Intensive mine development began in 1927, and copper production 
at the Roan Antelope began in 1931. Rhokana (Nkana) made its 
first production in December 1931, and Mufuhra in 1933. Up to the 
end of 1934 the district produced 356,300 tons of copper from 
12,276,385 tons of ore. 

The eventual tonnages of the Rhodesian mines are unknown, but 
they have been developed sufficiently to show that they are gigantic 
long-life deposits. The ore reserves and grades that have been officially 
announced (1935) are given in Table 14. 

TABLE 14 a 
ORE RESERVES AND GRADE OF RHODESIAN DEPOSITS 



Deposit 


Group 


Approximate 
Reserves 

(tons) 


Percentage 
of 
Copper 


Roan Antelope 
Nkana 
N'Changa 
N'Changa Extension 
Chingola 
Mufuhra 
Chambishi 
Baluba 


Roan 
Rhokana 
do . 
. do 
do . 
Mufuhra 
. do . 
do 


104,000,000 
127,000,000 
95,280,000 
46,500,000 
2,000,000 
116,000,000 
25,000,000 
21,000,000 


3.43 
4.0 
3.53 
6 9 
7 
4.41 
3 46 
3 47 



Bateman, A M , op cit 



450 PRODUCTION OF COPPER 

At present 10 there are four large companies operating in Northern 
Rhodesia. 

1. Roan Antelope Copper Mines, Ltd. The ore body m the Roan 
Antelope area is over 3% miles long and has a maximum depth of 
about % mile. Up to the end of 1938 all ore mined was from above 
the 820-foot level, and this ore was hoisted through the Beatty shaft. 
The Storke shaft, 2644 feet deep, will handle ore from below the 
820-foot level; this shaft is about 1% miles west of the Beatty shaft 
and is located near the center of the ore body. 

2. Mufuhra Copper Mines, formed in 1930, controls the mining 
rights on the Mufulira, Chambishi, and Baluba areas. 

3. Rhokana Corporation, Ltd , controls the Mmdola section and the 
Nkana section; the company operates a concentrator and smelter and 
also an electrolytic copper refinery and a cobalt segregation plant. 

4. N'Changa Consolidated Copper Alines, Ltd., was formed in 
March, 1937, to acquire the mining rights in four areas Chingola, 
N'Changa, Mimbula, and Kakosa. 



SOUTH AMERICA 

The two principal copper producing countries in South America are 
Chile and Peru, although smaller amounts have been produced in 
Bolivia, Argentina, and Venezuela. The copper production of Chile 
and Peru is given in Tables 15 and 17. 

Chile. Chile has been an important producer of copper since 1800, 
and this country has always contributed a substantial part of the 
world's total (Fig 2; Table 15). 

1 Chuquicamata, the mine of the Chile Copper Company, is 
located at Chuquicamata in the province of Antofagasta. This is the 
largest copper mine in the world, and its ore reserves are the greatest 
Utah Copper approaches Chuquicamata most nearly in tonnage of ore 
reserves, but the Chuquicamata ore is more than twice as rich as 
Utah's. The ore mined at Chuquicamata has been principally oxide ore 
up to the present and the copper has been won by leaching and electro- 
deposition. The Chile Copper Company is a subsidiary of the Ana- 
conda Copper Mining Company. 

2. Andes. The Andes Copper Mining Company, also a subsidiary 
of Anaconda, operates the mine at Potrerillos in the province of Ata- 
cama. This is the " youngest " of the porphyries from a production 

10 The Mineral Industry during 1938, Vol. 47, p. 152, McGraw-Hill Book Co , 
New York. 



PERU 



451 



TABLE 15* 
COPPER PRODUCTION OF CHILE 







Average 


Per Cent 


Per Cent 


Year or 


Production 


Yearly 


of South 


of World 


Period 


(short tons) 


Production 


American 


Total 






(short tons) 


Total 




1871-1880 


513,744 


51,374 


93.1 


36 08 


1881-1885 


226,402 


45,280 


85 7 


20 47 


1886-1890 


163,279 


32,656 


82.5 


11 81 


1901-1910 


363,264 


36,326 


67.7 


4.76 


1911-1920 


754,734 


75,473 


62.0 


6.20 


1921-1925 


827,799 


165,560 


76 


14 00 


1927 


264,242 


264,242 


81 2 


15.78 


1929 6 


350,000 


350,000 


83.8 


16.5 


1933 6 


180,000 


180,000 


86 


15.7 


1937 6 


455,000 


455,000 


91 5 


18.2 



a U S. Bur. Mines Econ Paper No 1. 

6 The Mineral Industry during 1938, Vol 47. 



standpoint. Andes maintains a concentrator and smelter for sulfide 
ores and a leaching plant for oxidized ores. 

3. Bradcn. The Temente mine of the Braden Copper Company is 
located at Sewell in the province of O'Higgms; Braden is a subsidiary 
of the Kennecott Copper Corporation. The ore mined has been prin- 
cipally sulfides which are concentrated and smelted at the smelter 
at Calctones. 

Table 16 gives the production of these three mines for 1936, 1937, 
and 1938 During this time these mines accounted for 91 to 92 per 
cent of the total Chilean production. 

Peru. The Cerro de Pasco Copper Corporation is the principal 
copper producer in Peru; the copper ore comes from two districts 
Cerro de Pasco and Morococha which lie about 70 miles apart. 
These mines are located in the high Sierra of central Peru in the 
Department of Junin. These deposits differ from those of most 
other important copper districts in that they include large amounts 
of lead and zinc. The lead-zinc ores, however, are mined separately. 

The copper production of Peru is given in Table 17. During 1936, 
1937, and 1938 the Cerro de Pasco Corporation produced respectively 
35,741, 37,547, and 39,230 short tons of copper; 11 these figures repre- 
sented, in turn, 97.4, 95.0, and 97.9 per cent of the total Peruvian 
production. 

11 Mineral Industry during 1938, Vol 47 



452 



PRODUCTION OF COPPER 



TABLE 16 
PRODUCTION OF COPPER FROM THE THREE LARGE CHILEAN COPPER MINES 





1936 


Production 
(short tons) 


Per Cent of 
Chilean Total 


Per Cent of 
World Total 


Chuquicamata 
Braden 
Andes 


123,433 
102,044 
30,027 


44.7 
36.2 
10.6 


7.60 
5.51 
1.62 




1937 


Chuquicamata 
Braden 
Andes 


200,402 
159,085 
60,478 


44.0 
34.9 
13.3 


8.00 
6.35 
2.42 




1938 


Chuquicamata 
Braden 
Andes 


163,213 
132,034 
60,880 


42.2 
34.2 
15 6 


7.48 
6.05 
2.76 



1 The Mineral Industry during 1938, Vol 47, McGraw-Hill Book Co , New York. 

TABLE 17 a 
COPPER PRODUCTION OF PERU 







Average 


Per Cent 


Per Cent 


Year or 


Production 


Yearly 


of South 


of World 


Period 


(short tons) 


Production 


American 


Total 






(short tons) 


Total 




1871-1880 


12,544 


1,254 


2.27 


0.88 


1881-1885 


2,285 


457 


0.88 


0.21 


1886-1890 


896 


179 


0.42 


0.06 


1901-1910 


145,945 


14,595 


27.3 


1.92 


1911-1920 


384,903 


38,490 


31.7 


3.17 


1921-1925 


203,655 


40,731 


18.7 


3.45 


1927 


52,438 


52,438 


16.1 


3.13 


1929 6 


60,000 


60,000 


14.4 


2.8 


1 1933 6 


27,400 


27,400 


13.1 


2.3 


1937 6 


39,400 


39,400 


7.9 


1.6 



U S Bur Mines Econ Paper No 1. 

6 The Mineral Industry during 1938, Vol 47. 



NORTH AMERICA 

The Continent of North America has long been the leading pro- 
ducer of copper in the world, as may be seen from the data in Table 18. 



MEXICO 



453 



We shall consider this continent in more detail than has been devoted 
to the other continents. 

The principal producer on the North American Continent is the 
United States, with Canada second in importance; both of these rank 
with the world's leading copper producing countries (Fig. 2). Mexico 
also produces a considerable amount of copper, and smaller amounts 
come from Cuba and Newfoundland. 

Copper production in North America began with Cuba about 1820. 
The United States production began about 1850; and Mexico, Canada, 
and Newfoundland began to produce copper about 1880. As Table 18 
shows, the importance of North America as a copper producer grew 
rapidly after 1870. In the period 1900-1920 the North American 
Continent produced more than two-thirds of the world's copper, but 
since then its rank has declined, largely for two reasons the increase 
in African production and the decrease in production from the United 
States. 

TABLE 18 
COPPER PRODUCTION OF NORTH AMERICA 







Average 


Per Cent 


Year or 


Production 


Yearly 


of World 


Period 


(short tons) 


Production 


Total 






(short tons) 




1871-1880 


221,245 


22,124 


15.54 


1881-1885 


308,220 


61,644 


27.88 


1886-1890 


562,684 


112,537 


40.69 


1901-1910 


5,127,971 


512,797 


67.23 


1911-1920 


8,219,061 


821,906 


67 44 


1921-1925 


3,573,679 


714,736 


60.46 


1927 


997,415 


997,415 


59.71 


1929 6 


1,250,000 


1,250,000 


58 9 


1933 6 


433,000 


433,000 


37.9 


1937 6 


1,161,000 


1,161,000 


46 5 



* U. S. Bur. Mines Econ Paper No 1 
6 Mineral Industry during 1938, Vol 47 

MEXICO 

Mexico has been a copper producer since about 1880, and for the past 
50 years has accounted for about 3 to 4 per cent of the world's supply 
(Table 19). 

There are three principal copper-producing districts in Mexico 
Boleo in Baja California, Nacozari in the State of Sonora, and Cananea 
also in Sonora. 



454 



PRODUCTION OF COPPER 



TABLE 19 
COPPER PRODUCTION OF MEXICO 







Average 


Per Cent 


Per Cent 


Year or 


Production 


Yearly 


of North 


of World 


Period 


(short tons) 


Production 


American 


Total 






(short tons) 


Total 




1881-1885 


2,116 


423 


71 


0.19 


1886-1890 


14,752 


2,950 


2.8 


1.07 


1901-1910 


554,118 


55,412 


12.9 


7.26 


1911-1920 


528,238 


52,824 


7.4 


4.33 


1921-1925 


216,082 


43,216 


7.0 


3.66 


1927 


63,760 


63,760 


7.5 


3.81 


1929 & 


86,700 


86,700 


6.9 


4.1 


1933 6 


43,700 


43,700 


10.1 


3.8 


1937 6 


51,500 


51,500 


4.4 


2.1 



U S Bur Mines Econ Paper No 1 
6 Mineral Industry during 1937, Vol 47 

Cananea. The Cananea district is the largest single producer in 
Mexico. The first copper mining at Cananea of which there is an 
authentic record was in 1881, although the district is reputed to have 
been the scene of mining operations for hundreds of years previous to 
this. The Cananea ore body produces gold, silver, and molybdenum 
as well as copper. The mines are operated at present by the Cananea 
Consolidated Mining Company, which is a subsidiary of Anaconda. 

Nacozan. The district of Nacozari has also been known as a mining 
district for hundreds of years, but the first important production came 
soon after the Pilares mine was acquired by the Phelps Dodge Corpora- 
tion in 1897. The present operating company is the Moctezuma 
Copper Company, a subsidiary of Phelps Dodge. 

Boleo. The copper deposit at Boleo was discovered about 1868, 
and about 1885 the French house of Rothschild acquired the property 
and formed the Compagnie du Boleo to work it on a large scale. This 
company has operated the mines ever since. 



CANADA 

Although Canada has produced copper since 1880, it was not until 
about 1900 that it became a real factor in world production. Its im- 
portance has increased greatly in recent years (Table 20, and Figs. 2 
and 5). In 1933 Canada accounted for over one-third of the North 
American production and almost 13 per cent of the world production 
(Table 20). 



ONTARIO 



455 



Quebec. There are two very active copper districts in Quebec 
Rouyn and Eastern Quebec. 

Rouyn. 12 The Rouyn district was the scene of a rush in the fall of 
1922 as a result of a gold strike, although the Home deposit had been 
staked 2 years before. The possibilities of this deposit as a copper 
mine was amply demonstrated when the Noranda Mines, Ltd., which 
was drilling in the Home mine, cut 130 feet of solid sulfide ore con- 
taining $4.36 in gold to the ton and 8.23 per cent copper. The Home 
mine is really a copper-gold mine since the value of precious metal 
produced is about equal to the value of the copper. 

TABLE 20" 
COPPER PRODUCTION OF CANADA 







Average 


Per Cent 


Per Cent 


Year or 


Production 


Yearly 


of North 


of World 


Period 


(short tons) 


Production 


American 


Total 






(short tons) 


Total 




1871-1880 


112 


11 


0.06 


0.01 


1881-1885 


5,365 


1,073 


1.82 


0.49 


1886-1890 


12,111 


2,422 


2.16 


0.88 


1901-1910 


253,663 


25,366 


4 95 


3.33 


1911-1920 


444,394 


44,439 


5.42 


3.65 


1921-1925 


196,645 


39,329 


5.50 


3.33 


1927 


70,698 


70,698 


7.07 


4.22 


1929 & 


121,000 


121,000 


9.7 


5.7 


1933 6 


148,000 


148,000 


34.2 


12.9 


1937 6 


262,000 


262,000 


22.6 


10.5 



U. S. Bur Mines Econ. Paper No. 1 

6 The Mineral Industry during 1938, Vol 47. 

Eastern Quebec. Eastern Quebec was one of the earliest producing 
districts in Canada ; the presence of copper ore was known as early as 
1841, and production dates from about 1858. The production from 
this district is small, however, when compared with the other great 
Canadian copper districts. 

Ontario. The copper-nickel mines of the Sudbury district make the 
Province of Ontario the largest copper producer in Canada. The first 
production was in 1886, and since then the district has developed 
steadily; today it produces about 90 per cent of the world's nickel, and 
is likewise one of the world's leading copper producers. 

12 The discussion of this, as well as the other copper districts in North America, 
is taken largely from Gardner, E. D., Johnson, C. H., and Butler, B. S , Copper 
Mining in North America : U. S. Bur Mines Bull. 405. 



456 PRODUCTION OF COPPER 

The principal company in the district is the International Nickel 
Company, and its two largest mines are the Frood and the Creighton. 
The Falconbridge Nickel Mines, Ltd., also operates in the district. 
The ores of the district yield important amounts of gold, silver, and 
platinum metals as well as nickel and copper. 

Manitoba. The copper deposits of The Pas district have been de- 
veloped only recently there was very little production from them 
before 1930, but since then the production has been increasing rapidly. 



1 520 



440 r" -^-British Columbia 



-360 
280 
200 
= 120 
S 40 





-Manitoba 
Saskatchewan 



1890 1895 1900 1905 1910 1915 1920 1925 1930 1935 

[(Data from Butt 406, U S Bur Mines, and The Mineral Industry During 1938, 

McGraw-Hill Book Co , New York) 

FIG. 5. Canadian Copper Production. 

There are two important mines in The Pas district the Flin Flon 
mine, owned by the Hudson Bay Mining and Smelting Company, and 
the Sherntt-Gordon mine 30 miles east of Flm Flon, which is the 
property of Sherritt-Gordon Mines, Ltd. The Flin Flon operates on 
a copper-zinc deposit containing small amounts of gold and silver. 

Since about 1933 Flin Flon has produced some copper from Saskatche- 
wan; the Manitoba property is near the boundary between the two 
provinces. In Figure 5 the production figures for the two provinces are 
lumped together. 

British Columbia. British Columbia from about 1900 to 1930 was 
the largest copper-producing province in Canada, but today it ranks 
below the eastern provinces. The mines near Rossland were large 
producers, but today they are idle. The other important districts in 
British Columbia are Hidden Creek, Anyox, and Howe Sound 
(Britannia) . 

Figure 5 shows the production of the various Canadian provinces 
from 1885 to 1937. 

THE UNITED STATES 

The United States, the largest producer of copper in the world, has 
maintained its supremacy for many years (Table 21; Fig. 2). From 
1901 to 1927 this country has consistently produced more than half 



MICHIGAN 



457 



of the world's copper, but in recent years its production has declined 
to about 30 per cent (Table 21), although the United States is still the 
largest producing country. 



Arizona 

Montana 

Michigan 

Utah 

Nevada 




1845 1855 1865 1875 1885 1895 1905 1915 1925 1935 

(Data from Bull 405, U. S. Bur Mines, and The Mineral Industry During 19S8, 

McGraw-Hill Book Co., New York] 

Fio. 6. Production of Copper m the United States. 

Practically all the copper produced in the United States has come 
from the following states and territories Arizona, Michigan, Mon- 
tana, Nevada, Utah, New Mexico, Tennessee, California, and Alaska. 
We shall now consider them individually. Figure 6 is a graphic 
representation of the amount of copper produced by the various states 
since 1885, and Figure 7 also shows other production statistics for 
both states and districts. 

Michigan. Native copper was discovered in the Lake Superior 
region in the seventeenth century by explorers who passed through, but 
the first production of copper from a lode mine did net come until 
1844. After this the Michigan copper district (the Upper Peninsula) 
soon became the leading producer in the United States and main- 
tained this position for many years. 

The Lake copper produced from the Michigan native copper ores 
is a very pure product, and the fact that there is a special commercial 
name for this copper speaks for its importance in the world market. 



458 



PRODUCTION OF COPPER 



TABLE 21* 
COPPER PRODUCTION OF THE UNITED STATES 







Average 


Per Cent 


Per Cent 


Year or 


Production 


Yearly 


of North 


of World 


Period 


(short tons) 


Production 


American 


Total 






(short tons) 


Total 




1871-1880 


208,768 


20,877 


94 2 


14 66 


1881-1885 


294,337 


58,867 


95 5 


26 62 


1886-1890 


526,071 


105,214 


93 5 


38 04 


1901-1910 


4,281,714 


428,171 


83 5 


56 13 


1911-1920 


7,160,559 


716,056 


87 3 


58 75 


1921-1925 


3,099,996 


619,999 


86 8 


52 44 


1927 


847,419 


847,419 


84 8 


50 60 


1929 6 


1,027,000 


1,027,000 


82 1 


48 5 


1933 5 


234,000 


234,000 


53.9 


20.4 


1937 6 


835,000 


835,000 


71 7 


33 3 



U. S Bur Mines Econ Paper No 1. 
b Mineral Industry during 1938, Vol 47 

Before the advent of the electrolytic refining process, Lake copper 
was the purest metal available 

The two principal operators in the district are the Calumet and 
Hecla Consolidated Copper Company and the Copper Range Company. 

Montana. Practically all the copper mined in Montana has come 
from the mines of Butte in Silver Bow County. Butte was discovered 
as a gold camp in 1864 but copper began to be mined shortly after, 
and the importance of the copper increased until in 1887 the Butte 
production exceeded that of the Michigan mines. For a number of 
years Montana was the leading copper-producing state, but it has 
now been surpassed by both Arizona and Utah. The Butte district, 
however, has produced more copper than any other district in the 
United States. 

Utah. The high rank of Utah as a copper producer is due to a 
single mine the mine of the Utah Copper Company in Bingham Can- 
yon near Salt Lake City. This is the greatest of the North American 
porphyry copper mines and is the second largest copper mine in the 
world Chuquicamata alone exceeds it in copper reserves. As far 
as total production is concerned, the Bingham district is third among 
North American copper districts; Bingham, however, is a single large 
mine, whereas the two districts which have produced a greater 
quantity of copper (Butte and Michigan) both include several mines. 
The Utah ore contains only small amounts of gold and molybdenum, 
but the ore tonnage is so great that this mine ranks second in the 
United States in the production of gold and molybdenum. 



cale then tftt State* 



Copper River 
19H-0.9 



Sbasta County 
1897-0.6 



Lake Superior 
1845-8.6 



Production by Districts 

First figure below name date of first production 

Second figure below name -total production through 

1934 in billions of pounds 
Figure above name-rank of district on the basis of 
total production through 1934 




Total production, United States ^H U" TotaJ P , oduc i 00 , UMttd StatM 

nd Alaska 1845-4337 ^ ^HJL < *nd Alaska 1937 

25,314,391 short tons ^^^T ~~^* 834,661 short tons 

07.87* distributed as shown ^^ 96.19^ distributed as fiowr 

State production figures In rr-llllons of short tons 

Total production for United States for the 92 years 1645-1937 is approximately 30 tfrnes the 
1937 production. Widths of lines axe In a ratio of 30 to 1. 



(Data on State from Mineral Yearbook, 1998, U. 8. Bur. Jtftnet. Data on District* from Butt. 405, U, & Bur. Afinat) 

FIG. 7. Production of Copper by Eight Principal States and Alaska for 1846-1937 and for 1937. Production 

of Fourteen Principal Districts through 1934. 



OTHER STATES 459 

Arizona. Arizona is the largest copper-producing state in the Union, 
a position it has held for many years. Arizona, however, contains 
several large copper-mining districts, and thus differs from Montana, 
Utah, and Michigan, each of which contains only one important dis- 
trict. The Arizona copper districts are listed below. 

Morenci. The Humboldt and Clay ore bodies are the two most 
important deposits in the Morenci district; Phelps Dodge is the 
only operating company in the district. 

Warren (Bisbee). The most famous mine in the Warren (Bisbee) 
district is the Copper Queen; another is the Calumet and Arizona. 
Phelps Dodge Corporation and Shattuck-Denn are the operators in 
the Warren district. 

Globe-Miami. The principal production in the Globe-Miami dis- 
trict has come from two large mines the Miami and the Inspiration 
operated by the Miami Copper Company and the Inspiration Copper 
Company, respectively. 

Ray The Ray mines of the Nevada Consolidated Copper Com- 
pany are at Ray, Arizona. This property is now a subsidiary of the 
Phelps Dodge Corporation. 

Superior. The Silver King ore body in the Superior district is owned 
by the Magma Copper Company. 

New Mexico. New Mexico contains only one large copper mine 
the Chino mine at Santa Rita, which is one of the porphyry copper 
mines This property is a subsidiary of the Kennecott Copper 
Company. 

Nevada. Nevada also has one large porphyry ore body the mine 
of the Nevada Consolidated Copper Company at Ely. This is also 
a Kennecott subsidiary. 

Alaska. There were two important copper-producing districts in 
Alaska the Beatson mine on Latouche Island in Prince William 
Sound and the Kennecott mines in the Copper River district. Both 
of these have been worked out, and there is no longer any considerable 
copper production from Alaska. For a time (Figs. 6 and 7) these 
deposits, particularly Kennecott, produced a great deal of metal. 

Other States. Tennessee has been a producer of copper since about 
1847, the copper coming from the heavy sulfide (pyrrhotite) deposits 
in the Ducktown district. At present these ores are treated to recover 
sulfur (for sulfuric acid), iron, copper, and zinc the copper might 
almost be considered a byproduct. 

Most of the production of California has come from the Shasta 
County district (Fig. 7). One of the leading producers in the State 
at present is the Walker mine in Plumas County. 



460 PRODUCTION OF COPPER 

PRODUCTION TRENDS 

The following discussion is taken verbatim from a paper by 
Croston 13 published in July, 1937, and this, together with Tables 22, 
23, and 24, taken from the same article, gives an interesting account 
of the trend of copper production throughout the world. 

In the years preceding the Civil War, and up to 1869, the principal in- 
dividual copper producer of the world was the Mansfeld mine which has 
produced more or less regularly since the twelfth century. In that year 
Calumet and Hecla forged ahead, although its production was less than 6200 
tons of copper. By 1877 the reorganized mines oi Rio Tmto took premier 
position, with an output of slightly more than 27,000 tons, and maintained 
leadership until displaced by Anaconda in 1892 with an output of 37,500 
tons. The first mine to produce more than 50,000 tons a year was Ana- 
conda, in 1896, and with the exception of the years 1905-1907, when Calumet 
and Hecla again led, and 1908-1909, when the Copper Queen led, it main- 
tamed its position as the world's greatest copper mine until just before the 
depression [1929]. 

A study of the producers of a quarter of a century ago reveals that there 
were about 150 mines producing copper in substantial quantities, but only 
26 had outputs in excess of 10,000 tons of metal annually Of these only 
10 produced more than 20,000 tons and but two more than 50,000 tons, while 
there was but one mine capable of turning out 100,000 tons a year. The 
beginning of work on the porphyries brought in an era of large-scale low- 
cost mining, and the introduction of the flotation process and leaching made 
possible reasonably high recoveries at low cost. Today, through mergers, 
consolidations, and integration of the industry, a half dozen large units 
control the destinies of the copper trade of the world, with another half 
dozen smaller units sharing moht of the remainder. 

These companies, several of which can each produce about 500,000 tons 
of metal annually, have garnered the choicest ore reserves of the world and 
will continue to dominate world production without serious competition 
for many years to come. The smaller companies treating richer ores but 
with higher costs and slender reserves will have to operate under the um- 
brella of the giants. Improved mining methods, flotation, leaching and other 
processes have been a much greater boon to the large low-grade producers 
than to the smaller and richer mines. Usually the ore bodies of these 
smaller producers are not susceptible to the economies of such mining 
methods, and the improvements m recoveries and lowering of costs of newer 
treatment processes are either not advantageously employable or exert but 
a minor effect on production costs. 

Having the technical skill and the necessary finances, it appears probable 
that any new deposits of size will gravitate to the control of the present 

13 Croston, J J , Recent Trends in Copper Production; Ore Reserves and Costs* 
Am Inst Mm & Met. Eng Tech. Paper 826 (Mining Technology), July 1937. 



PRODUCTION TRENDS 461 

great producers, as the funds required for the large-scale development and 
equipment of a great copper deposit are prodigious. The copper industry 
has completed the same cycle observed in other great industries the con- 
centration of the business into fewer hands. There will always be a con- 
siderable number of small copper producers, but they will no longer have 
any considerable weight in the industry. 

The period since the World War has seen the rise of the British as im- 
portant factors in world copper production, and today the streams of copper 
flow from mine to market quite differently from the way they did ten years 
ago At that time American producers dominated world markets, Ameri- 
can-owned companies controlled most of the Latin-American production 
and in addition refined mobt of the rest of the world's copper. Now 
Katanga refines its own production in Belgium, the Rhodesmn output goes 
to England for treatment, Canadian production is refined \\ithm the 
Dominion, while other Continental refineries are handling business formerly 
done here. Even the Japanese are treating some of the Chilean output as 
as well as their own, and recently have contracted to handle the output of 
the Granby [British Columbia] concentrator. 

Within the past few years a tariff wall has been erected for the protection 
of domestic producers America no longer consumes more copper than all 
of Europe, although she may do so again later, for Europe has been using 
more and more copper per capita for some years Should the trend con- 
tinue, European-controlled mines will share in a large part of this business. 
Domestic producers apparently will operate primarily to supply domestic 
demands, and American-controlled foreign producers will sell in foreign 
markets in competition with European-controlled companies, or, when prices 
are high enough and demand sufficient, will supplement the needs of the 
domestic fabricators. 

The next decade will \\itness the inauguration of several new large pro- 
ducers. One company, N'Changa Consolidated Copper Mines, Ltd , has just 
been organized with a capital of about $25,000,000 to exploit some of the 
ore bodies owned by Rhokana. Others are to be anticipated in Northern 
Rhodesia, Belgian Congo, and Uganda, all controlled bv existing copper 
interests There is also the possibility of copper from African colonies of 
France and Portugal if developments are favorable. In Latin America, Ana- 
conda has the Santiago property m reserve, and there are the Rio Blanco 
and Ferrobamba deposits A number of properties are under development 
in Sweden, Finland, Serbia, Turkey, and elsewhere in Europe, which might 
in the aggregate turn out substantial tonnages of copper. In Canada a 
number of properties are nearly ready for production, on some plants have 
been built but not yet operated on account of the condition of the market, 
but soon Sherritt-Gordon, Waite-Amulet, Aldermac, Normetal and perhaps 
Coast Copper will be adding their quota. Here in the United States, the 
Mountain City property of Anaconda is already in production. Howe Sound 
is building a concentrator for its Chelan property, while Phelps Dodge has 
a large tonnage available at Morenci when times are propitious. Bagdad 



462 PRODUCTION OF COPPER 

may get into moderate sized production in time. These potential pro- 
ducers of the future will m the aggregate, together with the expected in- 
creases in output of large existing producers, more than offset any decline 
in output through the exhaustion of older properties, and serve to assure 
the world of adequate supplies of copper for a long time to come. While a 
considerable increase in the total consumption of copper is to be expected 
with the passing years, the rate of increase of production of virgin metal 
should gradually taper, and more of the demand be filled by secondary 
copper. It is probable that secondary copper, important as it is now, 
will play a vastly greater role in the future. 

Table 22 gives the output of the world's principal copper mines The 
figures were assembled from the annual reports of the individual companies, 
official statements or private communications, except where noted as un- 
official estimates, calculations from copper content of ores, concentrates or 
matte, etc. The latest available data cover the year 1935, and the com- 
parison is made with 1932, the low point of recent copper production. The 
years 1929 and 1930 cover the culmination of the recent boom, while 1912 
enables us to look back a quarter of a century. The present dominance in 
the industry of the newer producers will be noted. Not included in the 
table are a large number of copper mines that were in production in 1912, 
but have shut down permanently. 

The above quotation, including Tables 22, 23, and 24, gives a pic- 
ture of the important copper mines of the world and their copper 
reserves as the situation existed in 1935. Of course these figures are 
changing as time goes on, and they will continue to change in the 
future. A few items which indicate some of these changes are: 

1. The Chmo porphyry deposit at Santa Rita, New Mexico (which 
was idle in 1935 and is not listed in Table 22) , is now in production, and 
the new copper smelter at Hurley, New Mexico, which was blown in 
in 1939 is smelting the concentrates from the Chino ore. 

2. With respect to the Canadian producers mentioned in Mr. Cros- 
ton's discussion, following are the production figures for 1938. 14 

1938 OUTPUT 

PRODUCER (SHORT TONS) 

Aldermac 6,195 

Normetal 2,350 

Sherritt-Gordon 14,511 
Waite-Amulet 8,886 

3. The property of the Howe Sound Company on Lake Chelan in 
the State of Washington is now milling about 1200 tons of ore per 
day in its mill at Holden, Washington. 

14 The Mineral Industry during 1938, Vol. 47, McGraw-Hill Book Co., New York. 



PRODUCTION TRENDS 



463 



TABLE 22 
ANNUAL OUTPUT, IN SHORT TONS, OF WORLD'S LEADING COPPER MINES* 



Company 


1935 


1932 


1930 


1929 


1912 6 


Custom Smelters 












Amer Smelt & Ref Co 


204,452 


138,648 


440,784 


616,398 


66,000 


Amer Metal Co 


129,061 


107,704 


235,666 


245,856 




U. S Smelt., Ref & Mm. Co. 


1,671 


531 


2,986 


2,558 


10,576 


Groups. 












Anaconda c 


258,972 


139,339 


323,706 


495,285 


173,117 


Kennecott^ 


209,135 


82,536 


173,058 


250,567 


130,000 


Phelps-Dodge* 


88,438 


41,544 


56,479 


88,590 


141,723 


Mines* 












Chile' (Chuquicamata) 


131,994 


40,685 


89,596 


149,788 


0) 


Katanga 


118,698 


59,595 


153,165 


151,008 


2,664 


International NickeF (Ontario) 


116,505 


28,832 


54,872 


40,917 


11,000 


Braden' 1 (Chile) 


112,010 


49,871 


80,993 


88,163 


4,750 


Anaconda (Butte)' (Montana) 


77,371 


48,787 


98,617 


148,507 


148,487 


Utah* 


59,233 


30,006 


80,569 


148,313 


45,683 


Roan Antelope 1 (Rhodesia) 


56,753 


42,233 


1,133 


(7) 


(7) 


Rhokana* (Rhodesia) 


56,447 


54,408 


(7) 


(;) 


(7) 


Bor (Yugoslavia) 


42,990 


33,245 


26,966 


22,790 


8,250 


Noranda (Quebec) 


37,239 


31,507 


38,071 


25,813 


(7) 


Nevada Consolidated' 1 (Nevada) 


33,830 fc 


29,992 


70,990 


1.33,137 


62,757 


Calument & Hecla & subsidiaries (Mich ) 


33,336 


16,899 


58,199 


67,347 


65,006 


Cerro de Pasco (Peru) 


31,989 


22,910 


43,200 


49,993 


28,219 


Mufuhra* (Rhodesia) 


31,498 


(7) 


(;) 


(7) 


(7) 


Andes' (Chile) 


28,641 


11,619 


47,023 


81,332 


(7) 


Rio Tinto (Spam) 


(I) 


(0 


(0 


(I) 


44,716 


Mansfeld (Germany) 


25,890 


27,703 


23,268 


25,235 


22,600 


Hudson Bay (Manitoba) 


24,509 


21,079 


1,186 


(;) 


(7) 


Greene-Cananea' (Mexico) 


21,061 


18,410 


21,212 


29,413 


24,630 


Nippon (Japan) 


(n) 


19,518 


24,865 


19,225 


(n) 


Furukawa (Japan) 


19,010 


20,035 


18,536 


17,886 


14,763 


Mount Lyell (Tasmania) 


15,642 


12,271 


10,995 


8,739 


5,159 


Magma (Arizona) 


15,193 


10,853 


15,942 


19,128 


(7) 


Miami (Arizona) 


14,870 


7,907 


34,600 


29,421 


16,416 


Mitsubishi (Japan) 


14,672 


15,207 


14,040 


12,092 


11,287 


Sumitomo (Japan) 


14,475 


14,005 


18,895 


20,792 


10,024 


Outokumpu (Finland) 


12,180 P 


7,329 


5,680 


4,960 


(7) 


Granby (British Columbia) 


11,752 


19,324 


23,41b 


30,427 


14,553 


Fujita 9 (Japan) 


(;) 


10,370 


10,954 


10,139 


8,795 


Cyprus? 


13,000 


2,305 


4,500 


5,000 


(7) 


Northern Peru Mm & Smelt Co r 


(n) 


(n) 


10,701 


11,158 


(7) 


Messina 1 (Transvaal) 


10,656 


10,598 


9,751 


7,616 


700 


United Verde Extension (Arizona) 


9,985 


17,876 


22,783 


32,056 


(7) 


Tennessee 8 


8,964 


3,815 


7,190 


7,624 


6,626 


Boleo (Mexico) 


8,670 


11,487 


13,889 


12,903 


14,168 


Copper Range (Michigan) 


8,380 


6,094 


11,897 


12,099 


38,909 


Indian Copper (India) 


7,728 


4,962 


3,331 


1,831 


2,683' 


Matahambre (Cuba) 


7,672 


6,533 


17,299 


16,515 


2,500 U 


Howe Sound (British Columbia) 


7,214 


1,103 


22,633 


21,516 


7,150 


M'Zaita (Charos) (Chile) 


6,704 


(n) 


4,100 


3,000 


3,480 


Naltagua (Chile) 


5,926 


5,460 


4,339 


6,770 


1,225 


Sulitjelma (Norway) 


5,811 


(0 


(0 


) 


5,500 


Bohdens (Sweden) 


5,785 


3,805 


(;) 


(7) 


(7) 


Tocopilla (Chile) 


5,264" 


(n) 


5,207" 


(n) 


250 


Disputada de_Las Condea (Chile) 


5,016 W 


(n) 


3,990" 


(n) 


(n) 


Orkla (Norway) 


<l) 


(0 


(0 


(0 


(0 


Mountain City (Nevada) 


4,100 


(;) 


0) 


(7) 


(7) 


Tharsis (Spain) 


) 


(v) 


(0 


(0 


3,782 


Burma* > w (Burma) 


3,946 


5,871 


8,416 


(n) 


(n) 


Inspiration (Arizona) 


3,758 


8,512 


32,803 


53,654 


(7) 


Mother Lode (Alaska) 


3,333 


1,716 


4,823 


6,121 


(7) 


Buchans r (Newfoundland) 


3,259 


2,373 


1,054 


1,032 


(7) 


Cons Copper & Sulphur (Quebec) 


2,339 


2/46 


2,489 


2,177 


925 


Falconbndge (Ontario) 


2,515 


1,197 


656 


(7) 


(7) 



464 



PRODUCTION OF COPPER 



TABLE 22 (Continued) 



Company 


1935 


1932 


1930 


1929 


1912* 


Mines (continued) * 












Mt. Morgan* (Australia) 


1,668 


259 


0) 


O) 


10,394* 


Shattuck-Denn (Arizona) 


1,381 


(v) 


4,167 


6,368 


1,255 


Minor producers during 1935. 












Walker (California) 


822 


535 


7,888 


7,516 


(;) 


Huelva Copper* (Spain) 


(n) 


1,534 


(n) 


2,281 


(n) 


Pena Copper (Spam) 


343 


645 


1,008 


923 


888 


Sunshine (Idaho) 


673 


357 


198 


133 


0) 


Carlo ta 


635 


(n) 


(n) 


(n) 


(n) 


Shenandoah-Dives (Colorado) 


367 


879 


629 


315 


(;) 


Ohio (Utah) 


380 


(v) 


1,024 


1,108 


3,277 


Cons Mm & Smelt of Canada* 7 (B C ) 


319 


384 


7,064 


4,173 


1,457 


Roros (Norway) 


211 


<J) 


(0 


(0 


650 


Mason <fe Barry (Portugal) 


(n) 


257 


226 


218 


3,963 


Ouancos (Chile) 


(/) 


(0 


(0 


(/) 


(0 


Majden Pek 


(0 


(0 


(0 


(0 


(0 


Plakalnitza 




(0 




<J) 


(0 


Bagdad (Arizona) 


86* 


0) 


(;) 


0') 


0) 


Important producers idle during 1935 












Consolidated Coppermines" (Nevada) 




870 


16,306 


11,366 


1,919 


Otavi Mines,"' aa (S W Africa) 




2,646 


16,645 


13,889 


6,435 


Qumcy w (Michigan) 




(v) 


5,470 


2,230 


10,317 


Sherntt-Gordon*' (Manitoba) 




4,965 


(;) 


O) 


(;) 


Mazapil" (Mexico) 




(V) 


4,378 


6,239 


6,778 


Corocoro United^ (Bolivia) 




(/) 


1 ,654 


4,528 


2,058 


Calif ornia-Engels v (Arizona) 




(V) 


2,036 


5,650 


O) 


Poderosa" (Chile) 




577 


2,711 


2,905 


2,539 


Iron Cap (Christmas ) v (Arizona) 




107 


3,462 


2,209 


(;) 


Tezuitlan*' v (Mexico) 




(i/) 


1,781 


1,719 


(66) 


Gatico (Chile) 




(v) 


(n) 


(n) 


1,936 


Mmdouh (French Congo) 




(v) 


198 


639 


0) 


Equipped for production* 












Wai te- A xnulet (Quebec) 


(v) 


(v) 


0) 


(/) 


(;) 


Aldermac (Quebec) 


(V) 


(*) 


(v) 


0) 


O) 



Includes principal groups and custom smelters Does not include U S S.R. 

6 Figures for 1912 include production of individual companies since consolidated 

c Own mines and subsidiaries (Chile, Andes, Greene-Cananea) Custom ores (including Walker 
and Mountain City) and toll treatments 

d Own mines in Alaska, Utah, Nevada (Arizona and New Mexico shut down since April 1933 and 
October 1934, respectively) and Chile 

* Properties in United States only since September 1931 (when Moctezuma in Mexico was shut 
down) including Calumet & Anzona and New Cornelia, but not United Verde, recently acquired 

? Included in Anaconda total 

Sales, not production 

h Included in Kennecott total 

* Year ended June 30 following 
3 Not yet in operation. 

* Ely alone, previous years include Ray and Chino. 

* No recent figures available approximate relative rank shown by position 
m Sagano8eki, Hidachi and smaller properties, in 1912 Hidachi alone produced 8651 tons. 
n Not available 

Year ended September 30. 

* Unofficial estimate 
q Kosaka only. 

r Included in American Smelting & Refining Company total 

* Not including Ducktown, since acquired. 

* Old Cordoba output 

tt Estimate under old ownership 

v Estimated copper content of concentrate. 

w Estimated copper content of matte. 

* Year ended May 31 following for predecessor. 
v Shut down 

* Small expen mental production. 
oa Year ended March 31 following. 
66 Shut down during 1912. 



PRODUCTION TRENDS 



TABLE 23 

WORLD COPPER RESERVES INCLUDING COPPER CONTAINED IN IMPORTANT TONNAGES 
IN OTHER METAL DEPOSITS 



Company 


Date of 
Estimate 


Reported 
Ore Reserves 

(short tons) 


Average 
Copper 
Content 
(per cent) 


Total Metallic 
Content of 
Deposit 
(short tons 
of copper) 


United States: 










Anaconda (Butte) 


Oct 1, 1935 


75,000,000 a 


4 00 


3,000,000 


Bagdad Copper 


Dec 31, 1932 


48,000,000 


1 08 


518,400 


Cons Coppermmes 


Apr 20, 19,14 


35,000,000 


1 10 


385,000 


Gray Eagle 




1,045,000 


3 23 


33,754 


Inspiration Cons 


Oct 1, 1935 


69,010,770 


1 373 


947,518 


Miami Copper 


Jan 1, 1936 


84,624 869 


938 


794,163 


Mother Lode 


May 26, 1934 


i04,000 


11 59 


12,054 


Mountain City 


Oct 1, 1935 6 






100,000 


Kennecott Copper c 










Ely property 


July 5, 1935 


77,000,000 


1 22 


939,400 


Ray property 


July 5, 1935 


81,448,000 


1 51 


1,229,865 


Chmo property 


July 5, 1935 


170,839,000 


1.23 


2,101,320 


Bmgham property 


Dec 3, 1930 


640,000,000 


1 07 


6,848,000 


National Copper 


1929 


900,000 


2.3 


20,700 


Old Dominion 


1931 


2,000,000 


2.00 


40,000 


Phelps Dodge d 


1932 


388,146,550 


1.13 


4,386,056 


Seneca Copper 


Dec 31, 1930 


3,486,895 






Tennessee" 


Dec 31, 1935 


6,148,132 


1 60 


98,370 


United Verde Extension . 


Dec 31, 1935 


61,000 


7 00 


4,270 


Van Dyke 


Apr 31, 1931 


1,000,000 


5 00 


50,000 


Not reported Unofficial estimates 










Michigan 


1934 


100,000,000 


1.00 


1,000,000 


California, Montana, New 










Mexico, Arizona, Tennessee, 










etc 




70,000,000 


2 143 


1,500,000 


Canada and Newfoundland f 










Aldermac 


Apr 1, 1936 


1,743,760 


2 00 


34,875* 


British Columbia Nickel 


Feb 28, 1936 


1,042,200 


46 A 


4,794 


Demson Nickel 


1936 


741,240 


1 05* 


7,783 


Falconbndge Nickel 


Dec 31, 1935 


4,059,475 


91' 


36,941 


Granby Cons 


Dec 31, 1933 


13,449,900 


1 81 


243,443 


Great Falls 


Dec 31, 1935 


300,000 


1 00* 


3,000 


Howe Sound (Britannia) 


Dec 31, 1935 


10,000,000 


1 30 


130,000 


Hudson Bay 


Dec 31, 1935 


24,770,000 


2 10* 


520,170 


International Nickel 


Dec 31, 1935 


205,590,592 


2 00 m 


4,111,812 


Mandy 


1929 


98,000 


6 5 n 


6,370 


Noranda 


Dec 31, 1935 


31,029,000 


2 51 


779,082 


Normetal 


Jan 1, 1934 


700,000 


3 OOP 


21,000 


Ontario Nickel 


Apr 1, 1936 


116,000 


1 04 


1,206 


Sherntt-Gordon 


Dec 31, 1932 


4,783,175 


2 41 r 


115,419 


Sudbury Basin 


Dec 31, 1929' 


833,000 


2 80' 


23,533 


Sunloch 


1918* 


4,500 


3 00 


4,500 


Wai te- A mulct .... 


Dec 31, 1933 


955,445" 


4 50 


43,023 


Buchans 


Dec 31, 1934 


6,375,000 


1 50 V 


94,625 


Newfoundland 










Gull Lake 


1929 


1,300,000 


2.70 


35,100 


Mexico - w 










Greene-Cananea 


Oct. 1, 1935 






240,000* 


M^octezuma" 


1930 


3,500,000 


2.70 


94,500 













466 



PRODUCTION OF COPPER 



TABLE 23 (Continued) 



Company 


Date of 
Estimate 


Reported 
Ore Reserves 

(short tons) 


Average 
Copper 
Content 
(per cent) 


Total Metallic 
Content of 
Deposit 
(short tons 
of copper) 


Cuba: 










S A Mmaa de Matahambre 


1933 


939,000 


4 75 


41,603 


Nicaragua 










Tonopah Nicaragua 


1918" 


1,492,088 


5 059* 


75,485 


Veneuela 










South American 


1932 


400,000 


3 50 


140,000 


Peru aa 










Fundicion de Magistral 




3,000,000 


7 00 


210,000 


Ferrobamba 




6,778,000 W ' 


3 56 


241,450 


Chile cc 










Andes Copper 


Dec 31, 1935 


112,549,287 


1 47 


1,054,475 


Kennecott (Braden) 


Dec 31, 1935 


205,019,050 


2 18 


4,409,415 


Chile 


Dec 31, 1035 


9*8,931,000 


2 15 


20,187,023 


Rio Blanco dd 




42,000,000 


2 3 


966,000 


Santiago** 


Dec 31, 1921 


52,819,282 


2 11 


1,110,078 


Australasia 










Lake George 


1931 


2,880,000 


75" 


21,600 


Mount Elliott 


1929 


1,910,550 




100,936 


Mount Lyell 


Sept 30, 1930 


8,825,000 


1 W 


167,075 


Mount Morgan 


June 30, 19 }b 


7,500,652 


1 77 M 


132,868 


Wallaroo and Moonta 


1935 


618,000 


3 77 


23,302 


Asia " 










Burma 


June 30, 1930 


3,914,1*2 


87" 


34,053 


Indian 


Dec 31, 1935 


950,801 


3 19 


30,331 


Europe ** 










Huelva Copper 


mm 1933 


3,358,739 


1 10 


36,936 


Huelva (Pyrites de) 


tnm 1933 


6, 613,^60 


5S 


38,3(>0 


Imperial Chemical. 


mm 1933 


3,913,201 


1 04 


40,593 


Pena 


mm jq.j 3 


3,314,634 






Rio Tmto 


mm JQ33 


5,000,000 


1 40 


70,000 


St Gobam. 


mm 1933 


1,653,465 


1 39 


22,983 


San Platon. 


mtn 1933 


440,924 


3 50 


19,290 


Seville 


mm ^33 


5,511,550 






Tharsis 


mm 1933 


100,861,365 


75 


756,460 


Bohdens Gruvaktiebolag 


1934 


5,900,000 


2 00 


118,000 


Outokumpu 


nn 1933 


9,369,635 


4 00 


374,785 


International Nickel-Petsamo 


nn 1934 


5,511,550 


1 30 


71,650 


Foldal 


00 1933 


330,693 


1 63 


5,401 


Grong 


00 1933 


8,487,787 


2 21 


187,392 


Orkla 


1935 


20,000,000 


2 5 


500,000 


Roros 


00 1933 


771,617 


1 00 


7,716 


Sulitjelma 


1933 


5,070,911 


1 77 


89,622 


Mansfeld A G 


pp 1933 


59 




650,163 


Rammelsberg 


pp 1933 


1,615,986 


4 27 


69,005 


Stadtberger Kupferhutte 


PP 1933 


688,937 


1 6 


11,023 


Sontrau-Richelsdorf 


1936 


rr 




440,924 


Plakalnitza 


" 1933 


165,347 


2 50 


4,134 


Cyprus . . 




20,000,000" 


2.10 


420,000 


U.S.SR uu 










Developed or drilled 








6,302,347 


Probable 








4,146,449 


Possible 








7,421,412 


Total 




1,532,398,293 




17,789,468 



PRODUCTION TRENDS 



467 



TABLE 23 (Continued) 











Total Metallic 


Company 


Date of 
Estimate 


Reported 
Ore Reserves 
(short tons) 


Average 
Copper 
Content 


Content of 
Deposit 
(short tons 








(per cent) 


of copper) 


Africa vv 










Katanga 


Dec 31, 1934 


78,000,000 


6.50 


5,070,000 


Sud Katanga 


1934 








Tchmsenda Aline 




16,534,650 


4.00 


661,386 


Mokambo Mine 








338,614 


Mufulira 


June 30, 1936 








All properties 




160,390,000 


4.12 


6,608,068 


Rhodesia- Katanga 


Aug 1, 1936 








Kansanshi Mine 




11,000,000 


4 25 


467,379** 


llhokana 










All properties 


June 30, 1936 


269,454,435 


4 12 


11,101,523 


Roan Antelope 


June 30, 1936 


95,637,987 


3 43 


3,280,383 


Falcon Mines 


1933 


56,000 


7 3 


4,088 


Messina (Transvaal) 


June 30, 1936 


1,268,030 


2 231 


28,290 


Zambesia 










Kolembe Mine 


1932 


330,000 


4 50 


14,850 


a Boston News Bureau, Dec 16, 1935 " The engineering expert \\ho examined the Montana 



properties for the bankers gave it as his opinion that the Butte mines can produce 300,000,000 Ib of 
copper annually for (he next 20 years " The estimate of 75,000,000 tons is based on the assumption 
of a 4 per cent content, which is not official In the prospectus of the $55,000,000 issue of debentures, 
dated Get 1, 1935, Mountain City was credited with 100,000 tons of copper metal 

b Anaconda Copper Mining Co prospectus dated Oct 1, 1935 Materially increased since the 
date of estimate 

' Not including Alaskan or Chilean properties 

d Including Moctezuma in Mexico, but not United \erde, since acquired, which has substantial 
reserv es 

e Not including Ducktown, acquired m 1936 

S No data available on the reserves of Coast Copper, George Gold and Copper or Island Copper, 
all in British Columbia, or the long established Consolidated Copper and Sulphur Co , Ltd , in Quebec 

No recent data are available regarding Newfoundland properties although substantial amounts 
of copper are contained at the Tilt Co\e and other deposits. 

Plus 75 oz siher and 70c gold ((g) $.35/oz ) per ton 
h Plus 141 per cent nickel 

1 Plus SI per cent nickel and $3 in gold and platinum 
1 Plus 1 93 per cent nickel 

* Arbitrary assumption of 1 per cent Company reports 2 per cent combined copper and nickel 

1 Plus 3 SG per cent zinc, 1 28 oz silver and $2 SO gold (@ $35/oz ) per ton 

m Estnnated, curiently treating ore higher in copper, and with nearly 3 per cent nickeL 

n Plus 16 per cent zinc and $5 gold and silver (at prices then prevailing) per ton. 

Plus 5,605,515 oz of gold 

p Plus 13 5 per cent zinc and 4 3 oz silver per ton 

q Plus 87 per cent nickel, some platinum, gold and silver. 

r Plus 61 c gold and silver (at prices then prevailing) 

' No change as of Dec 31, 1935 

* Plus 5 8 per cent zinc, and minor amounts of lead, silver and gold. 
u Not including 300,000 tons of 11 52 per cent zinc ore 

v Plus 15 8 per cent zinc, 7 7 per cent lead, 3 60 oz silver and 05 oz gold per ton. 

No data available on Boleo, Inguaran, Mazapil, Tecolote, Tezuitlan or Triunfo. 

z Anaconda prospectus for $55,000,000 debentures, dated Oct 1, 1936, gives minimum reserves 
sufficient to produce a rate of 60,000,000 Ib for 8 years Probable reserves are believed by others to 
bring the total close to three-quarters of a million tons of metal 

v Included in Phelps Dodge total 

* Plus $1 54 gold (@ $35/oz ) per ton. 

ao No data available for Cerro de Pasco or Northern Peru 



468 PRODUCTION OF COPPER 

66 Not exploited since date of estimate A recent estimate is 2,013,000 tons of 3 98 per cent copper 
within 100 ft of surface, and an indeterminate tonnage of about 3.15 per cent copper below. 
No data available on Bolivian reserves (Corocoro). 

cc No data available on M'Zaita (Chagres), Chanaral, Disputada, Copiapo, Gatico Naltagua, 
Ouancos, Poderosa or Tocopilla. 

dd According to A H Rogers (January 1937) 

'* No exploitation since date of estimate. 

H Plus 12 94 per cent zinc, 7.5 per cent lead, some gold and silver. No exploitation since date of 
estimate 

00 Plus some gold and silver. 

** Plus 218 oz gold per ton 

tl No data available on reserves of Japanese producers Fujita, Furukawa, Mitsubishi, Nippon 
or Sumitomo. 

77 Plus 23 6 per cent lead, 14 5 per cent zinc, 18 2 oz silver per ton, some nickel 

** No data available on reserves of Bor or Majden-Pek 

mm Spanish reserves from " Cobre en Espana," Int Geol Congress, 1933 Huelva copper is credited 
with an additional 795,868 tons of probable ore same grade Pyrites de Huelva is credited with an 
equal amount of probable ore of the same copper content Imperial Chemical Industries, Ltd , has 
4,034,455 tons of probable ore containing 40,708 tons copper metal Rio Tinto has 166,786,000 tons 
positive pyrite and 77,162,000 tons of probable pyrite ore, in addition to the porphyry St. Gobain 
has probable reserves of 7,700,000 tons of pyrite. San Platon has an equal amount of probable ore of 
same copper content. Tharsis has 106,097,338 tons of probable ore of the same copper content. 

nn Finnish reserves from Saksela Int Geol Congress, 1933 

00 Norwegian reserves, except Orkla, from Foshe. Int Geol Congress, 1933. 

pp Mansfeld, Ramrnelsberg and Stadtberger reserves from Fulda Int Geol Congress, 1933. 

qq Mansfeld reserves are given as 23 sq km , or somewhat over 20,000,000 tons of copper ore. Of 
this 8 km are developed, and 15 still are undeveloped In the 700 years of its existence 117 sq km 
have been worked out Rammelsberg has an additional 38,581 short tons of copper metal in probable 
reserves 

rr According to C W Wright- Spec Sup No 3, Mineral Trade Notes, U. S Bur Mines (Sept 19, 
1936) This company under management of Mansfeld and under heavy German subsidy, has some 
480 workers, has sunk a number of shafts and drill holes and developed this estimated tonnage 

** According to Boncev Int Geol Congress, 1933 

" Skounatissa mine only Mavrovonm deposit now largest producer, but no data available on 
tonnage 

Mu Based on figures of Riddell and Jermain Russian Copper Eng and Mm Jour. (Feb 1935) 

** No data available on reserves of Mindouli (French Congo); Bembe (Portuguese West Africa); 
Kafue (Northern Rhodesia), Umkondo (Southern Rhodesia), Tsumeb (Southwest Africa); South 
African Copper (Cape Province) or Northern Transvaal (Messina) Copper in Transvaal 

ww Including Mokambo reserves the total is more than 1,000,000 tons of metal, although Mokambo 
tonnage is not stated 

xx Plua 220,000 oz. of gold, or 70c per ton (gold @ $35/oz.). 



PRODUCTION TRENDS 



469 



TABLE 24 
SUMMARY OF WORLD'S REPORTED RESERVES OF COPPER IN SHORT TONS 





Tonnage of Ore 


Metal 




Region 


Officially 
Reported 
and with 
Grade 


Unofficially 
Estimated 
and with 
Grade 


Officially 
Reported, 
Unofficial 
Estimate 


Officially 
Reported, 
No Grade 
Stated 


Content 
Officially 
Reported 
but Ore 
Tonnage 
and Grade 


Tonnage of 
Metallic 
Copper in 
Deposits 








of Grade 




not Stated 




United States 


1,605,327,321 










18,408,870 






170,000,000 








2,500,000 










3,486,895 
















3,100,000 


3,100,000 


Total ... . 












24,008,870 


Canada . 


102,000,695 










2,101,864 








205,890,592 






4,114,812 


Total 












6,216,676 


Mexico, Cuba, Central 














and South America 


1,364,427,707 










28,989,029 






3,000,000 








210,000 












240,000 


240,000 


Total 












29,439,029 


Australasia 


21,746,202 










446,381 


Asia 


4,864,983 










64,384 


Europe (not including 














USSR) 


199,754,017 










2,843,360 










8,826,184 
















1,091,087 


1,091,087 


Total 












3,934,447 


Africa 


632,671,102 










27,235,967 












338,614 


338,614 


Total 












27,574,581 


USSR 














Developed or drilled 










6,302,347 


6,302,347 


Probable 










4,146,449 




Possible 










7,421,412 




Total 










17,879,468 




World Total by classes 


3,930,792,027 










80,089,855 


(not including 




173,000,000 








2,710,000 


U.SSR) 






205,890,592 






4,114,812 










12,313,079 
















4,769,701 


4,769,701 


Grand Total not in- 












91,684,368 


cluding USSR. 














Grand Total including 












97,986,715 


USSR 














Grand Total including 












109,554,576 


USSR probable 














and possible 















470 PRODUCTION OF COPPER 

SMELTERS AND REFINERIES IN THE WESTERN HEMISPHERE 

In our discussions of metallurgical processes in previous chapters 
we have mentioned many smelters and refineries. At this point 
we shall briefly summarize the essential facts about the copper treat- 
ment plants in the Americas. Figure 8 is a schematic drawing which 
indicates the location of these plants, and Table 25 gives the essential 
data concerning them. 

Smelters and refineries represent large capital outlays, and they are 
not erected unless there is a reasonable assurance that there will be 
an adequate supply of raw material for their operation. Copper 
smelting and refining is not practicable on a small scale, and smaller 
mines generally concentrate their ore and then ship the concentrate to 
one of the large established smelters. 

A few words of explanation are necessary in connection with Figure 8 
and Table 25. The map in Figure 8 does not show all the smelters 
and refineries in the Western Hemisphere, but it does include all the 
important ones; most of these have been referred to in previous 
chapters. There have been copper smelters which were important 
in the past but are no longer operating these have been omitted. 
All in all, however, the plants shown on the map smelt and refine the 
great bulk of American primary copper, as well as a large proportion 
of secondary copper. Strictly secondary copper plants are not shown. 
Another smelter is now (1941) under construction at the Morenci 
property of the Phelps Dodge Corporation. 

Only one native copper smelter is shown in Michigan; there are 
others (Table 25), but all are in the same district. Two are operating 
at the present time. 

The copper refineries on the eastern seaboard of the United States 
and Canada have been indicated as electrolytic refineries to emphasize 
the difference between them and the smelters located near the pro- 
ducing mines. Although it is probably true that the most important 
function of these plants is the electrolytic refining of crude copper, 
some of them include complete smelting equipment (Table 25) and 
can handle copper ores, concentrates, and scrap of various kinds, as 
well as blister copper. 



SMELTERS AND REFINERIES IN WESTERN HEMISPHERE 471 



Hudson Bay f N on 

L /^S 1 

J FUn Ho^nltob. ^ 



A.C M Anaconda Copper Mining C 

f o Phelps Dodge Corp. 

k a AR Amer Smelting & Ret Co. 

KENN Kennecott Copper Corp. 

i N International Nickel Co 




FIG. 8. The Principal Copper Metallurgical Plants in the Americas. 



472 



PRODUCTION OF COPPER 





fc 





illllll III I, 

Jo O 






9 ^ 

it 



1 i 

M 



g 8 




g 8 

"5 00 




3 M | pj 

sLf.l 

>, g>| a5 

a S W >, 

Una 



I 




SMELTERS AND REFINERIES IN WESTERN HEMISPHERE 473 



o $ v "5 

^ifl 



IS 



a . e 
l! 



a ao 

IK 

8 8 |. 
1ft 

. g 03 



W ~* J I? ? <K >v w t 

O O 



S S 



: ! 



i S 

i-i 



U 

I 

5 



2 51 



I-sj: 

*3 a a _ 

s ;i i 

e I o 2 1 

a 1 -S c 

1 11 !. 



li 



^ e 

< O *- : 



I s 

^l- 8 , 
ISI| 

15 I I 
6 

g - 
o o 



l&l 

** >* 



gi 



^ 5 f 
2 s I 

C< O <N ( 



I 



fe > "S > 

* a c 

So > o 

K 5 

rx <N PH rf< 



fc > 

> a 

8 



11 

'" 



E? a 




21 



I 

s 

I 
S 



g 



II 

ffi 



Cale 



c 

i 
i 



I 



ii 



i 



> 

s 



I I 



474 



PRODUCTION OF COPPER 





rom M 



5M 

ill 

d a 3 

m 

8 go 

&i 

a e " 



ine and 
Rouyn 



If g a 

a * P 

I ^ 3 

fl w 

c3 0> TJ 

JgSI 

- * 2 



12- 
|||1 

o s ll! s 
gi^-a go g s 
o o 



al 




l 



II 

tl 



O e8 

li 

v 
a 



3 
S 



I 

w 
S 



Equipm 



3 8 



(N CO i-* ^H 



is 

1 s 

s 

2 8 



2 " 





1 



la 



Q. 

a 



1 



p S 

6 



q 

>>r 



! 



a 




k 

a 

a 



SMELTERS AND REFINERIES IN WESTERN HEMISPHERE 475 




-s 
a g 



3 



I 
a 

2 g 






w ^ 



i" 









- 



















X 



a 




c 










0) 


1^ 


o3 




4) 










c 


s. 


g 




G 













o. 
g 





G 


8 










o. 


o 


c3 


o 
"3 
-3 


1 

a 




-8 






8 





a 




o 
o 










(D 


S 




a 


<u 


3 








I 


1 


a 


i 


03 


1 








G 









ao 






o 


i 


MELTERS 


) tons of conce 


;rate per meltii 


>> 

oS 
T3 
O 

1 




000,000 pound 


jopper per yeai 


o 


)f copper per > 


QQ 


g 








f 









O 


a 




1 








s 


? 


O 


b 




b 








c 


c 









o 













w 


rt 


03 




c3 






o3 


CO 







g 


1 


| 






1 


" 


<1 


fe 


<-c 


flj 


+H 














bC 


> 


bO 















G 





g 









s 




<N 




CO 








10 








4, 






1 












o 
















1 

K 




1 






6 




! 






oS 


1 


o 




I 




1 


I 




"o 

1 


I 




I 




1 

a 

o 


G 

a 


1 


a 
5 




6 













a 






o 










1 


o 














$ 


"2 


8 







1" 








1 


^ 


03 


s 




B 


i 






1 


" 


1 


* 




1 








1 










i 


w 






S. 




| 






3 








1 




I 






i 'S a ' 

- 1 a o 



ter. 



8 <H -S 8 G 

l!l!i 



s 8 

II s 1 1 S a 

: g | s -as s 



o ^ 



I 

S 



OQ flj 

"e 2. 



Hil 
III^ 

111! 

8 ' 81 

<N CO O 



as- 



a o 



4 



O 



I' 



|| 

II 



476 



PRODUCTION OF COPPER 



Source of Raw 
Matenal 



-s 

c 



H 
?H 
J 



I 

1 



f>> 
i 

1 

la 



. M > 3 -3 \ 

If III 



1 
1 



ij 

Sll 

IM 

EfJ 

|1 

CQ 



I 
3 

JJ 

t 00 

s! 
If 



w >< 

I 1 





i. 

8 



I 



& 

R 



o - 

I 



8 
"8 

3 



S 
5 



S fe 8 j 
<N fi, o S, 




3 



> ^ ^ "O 

I -J* fl> C _ 

lilllf 

> co W 



If || 
"111 



r-( Tt< CO 








It 

la 

02 9 




1 

s 



Q b- 

&si 



s ^ 

111 



*b 




SMELTERS AND REFINERIES IN WESTERN HEMISPHERE 477 



1 
I | 



o 

I 



6 

US! 

' 



8 ' * ! 

rH CO i 



Jill- 



l-33-a 



II 
Hi 

isf 

2 Z g 



^|1 

Ill-s 

13|| 
=S S S 



a 



w 



I' 




' 8 2 | a c 
i *-, S J.g ~ 

fiilailll 

MOQ^C.Sv^aQC 



1! 

I! 



3 ^ 
o g 



sr 




13 



i- 

L-3 



I 

r 



478 



PRODUCTION OF COPPER 




I 

o 

15 



Eq 



Operati 
Compa 



1 

c o5 

J^I 
ell 



2 

fi 
II 



THE PRICE OF COPPER 



479 



Ammonia Leaching Plants. The only operating ammonia leaching 
plants in the Western Hemisphere are located in Michigan and are 
operated by Calumet and Hecla. They are used to leach finely divided 
native copper from conglomerate ores and reclaimed tailing from 
older operations. Ammonia leaching is operated in connection with 
gravity and flotation concentrations. 

THE PRICE OF COPPER 

Copper prices are quoted in cents per pound on the New York market, 
and in pounds sterling per long ton on the London Metal Exchange. 
The price fluctuates in a series of peaks and depressions, not only in 
annual average from year to year, but from month to month and 



220 

200 

=180 

o 

^160 

4 120 
100 

^ 80 

1 60 

2 40 
20 





High * 




47% 1 
43V 8 
39 | 
34% J 
30V 8 S 

ol 

26 *> 



13 T 

<u 

873 I 

4V' 3 I 



coa5Or-tQoo*0-m co r co o> o *-* ojco * 
r^r^-cocooocococo co oo oo oo CT> 05 o>o> en 

(From The Mineral Industry During 1938, AfcGraw-HiU Book Co., New York) 

FIG. 9. London Copper Prices. 

even from day to day. Prices are quoted on " spot copper " for im- 
mediate delivery, and on " copper futures " for delivery at some 
stated time in the near future (1 month, 3 months, etc.). 

Figure 9 is a graph of the price of copper on the London Exchange 
from 1780 to 1938 this curve shows the high and low quotations 
for each year. This figure shows that the all-time high came during 
the Napoleonic wars when the price reached 200 per long ton; the 
lowest price was during the depression in 1932 and 1933 when copper was 
quoted at less than 30 per long ton almost a hundred-fold decrease 
from the highest price. In spite of the violent fluctuations in price, 



480 



PRODUCTION OF COPPER 



this graph shows that the price of copper has gradually decreased 
from 1800 to the present time. 

Just before 1890 there was an effort (" The Secretan Syndicate ") 
to " corner " the world's copper market, and in 1930 the United States 
producers attempted to peg the price of copper at 18 cents per pound. 



36 
32 
28 
24 
20 

16 
12 



High 




~ 1900 1905 1910 1915 1920 1925 1930 1935 

(From The Mineral Industry During 1938, McGraw-Hill Book Co., New York) 

FIG. 10. United States Copper Prices. 

Neither of these efforts had any long-time effect on the price of 
copper in fact both were followed by sharp drops in price (Fig 9) . 

The all-time high copper price in the United States came in July, 
1864, when the average price was 55 cents per pound; the lowest price, 
4.775 cents per pound was quoted in January and February, 1933. 
Figure 10 shows the high, low, and average New York prices during 
the period 1900-1938; note that these prices show practically the same 
fluctuations as the London prices during the corresponding period 
(Fig. 9). 

The graphs in Figure 11 show the fluctuation of the average yearly 
New York prices the lower curve gives the actual market prices in 



COST OF PRODUCING COPPER 



481 



cents per pound, and the upper curve gives the " intrinsic " price of 
copper as corrected for the purchasing power of United States money. 
This curve shows smaller fluctuations than the lower curve and empha- 
sizes the general downward trend of copper prices. 



20 

115 



10 

v> 

I 5 




(A) (B) (C) (D) (E) (F) (G) (H) 



Relative Price of Copper 
(Index of Same Year) 
(Using Sauerbeck-Statist Index) 




(I) 




L J I I 



I I 



10 2 

I 

5 



1880 1885 1890 1895 1900 1905 1910 1915 1920 1925 1930 1935 

(Croston, Am. Inst Mm and Met Eng Tech Paper 826, Mm Tech., July 19S7) 

FIG. 11. United States Copper Prices. 

COST OF PRODUCING COPPER 

When copper is quoted at 12 cents per pound in New York that 
means that the buyer will pay 12 cents per pound for electrolytically 
refined copper cast into standard wirebars, cakes, billets, etc , and 
laid down in New York or vicinity. Certainly then a pound of copper 
disseminated throughout a 2 per cent ore buried 1000 feet below the 
surface of an Arizona mountain does not represent 12 cents worth of 
value to its owner. The total cost of mining, smelting, refining, and 
transporting metal must be less than the market price before there is 
any profit in the operation. High-cost operations must necessarily 
cease when the price of copper drops. 

Let us briefly mention a few of the factors which govern the cost of 
copper production: 

1. Size of ore deposit. Large-scale operations always show lower 
costs per ton of ore mined than smaller mining operations, and unless 



482 PRODUCTION OF COPPER 

the small deposit is unusually rich, the large deposit will generally show 
a lower mining cost per pound of copper. 

2. Nature of the ore deposit. The structure of the ore body deter- 
mines the mining method to be employed Disseminated ore deposits 
which can be mined by bulk methods such as open-cut or block-cav